From a1e8878ade6a4fafdbf9840d62836c515d548051 Mon Sep 17 00:00:00 2001 From: katriendg Date: Wed, 29 Apr 2026 14:07:21 +0200 Subject: [PATCH 1/9] feat(agents): add shared RPI context discipline instructions and lean response formats MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit - introduce hve-core/rpi-context-discipline.instructions.md for parent RPI agents - reference context discipline from rpi-agent, task-researcher, planner, implementor, reviewer - tighten subagent response formats to executive summaries with full detail on disk - update hve-core, hve-core-all, and project-planning collections plus generated plugins ๐Ÿงน - Generated by Copilot --- .github/agents/hve-core/rpi-agent.agent.md | 4 +++ .../implementation-validator.agent.md | 20 ++++++------ .../subagents/plan-validator.agent.md | 20 ++++++------ .../subagents/researcher-subagent.agent.md | 15 +++++---- .../hve-core/subagents/rpi-validator.agent.md | 15 +++++---- .../agents/hve-core/task-implementor.agent.md | 4 +++ .github/agents/hve-core/task-planner.agent.md | 4 +++ .../agents/hve-core/task-researcher.agent.md | 4 +++ .../agents/hve-core/task-reviewer.agent.md | 4 +++ .../rpi-context-discipline.instructions.md | 31 +++++++++++++++++++ collections/hve-core-all.collection.md | 1 + collections/hve-core-all.collection.yml | 2 ++ collections/hve-core.collection.md | 19 ++++++------ collections/hve-core.collection.yml | 2 ++ collections/project-planning.collection.md | 1 + collections/project-planning.collection.yml | 2 ++ plugins/hve-core-all/README.md | 1 + .../rpi-context-discipline.instructions.md | 1 + plugins/hve-core/README.md | 19 ++++++------ .../rpi-context-discipline.instructions.md | 1 + plugins/project-planning/README.md | 1 + .../rpi-context-discipline.instructions.md | 1 + 22 files changed, 122 insertions(+), 50 deletions(-) create mode 100644 .github/instructions/hve-core/rpi-context-discipline.instructions.md create mode 120000 plugins/hve-core-all/instructions/hve-core/rpi-context-discipline.instructions.md create mode 120000 plugins/hve-core/instructions/hve-core/rpi-context-discipline.instructions.md create mode 120000 plugins/project-planning/instructions/hve-core/rpi-context-discipline.instructions.md diff --git a/.github/agents/hve-core/rpi-agent.agent.md b/.github/agents/hve-core/rpi-agent.agent.md index db5b3b67c..49b2d588d 100644 --- a/.github/agents/hve-core/rpi-agent.agent.md +++ b/.github/agents/hve-core/rpi-agent.agent.md @@ -91,6 +91,10 @@ When a task requires subagents but neither `runSubagent` nor `task` tools are av Treat the phase guidance below as operating defaults rather than ceremony. Delegate only when it materially improves the outcome. +## Context Discipline + +Follow all rules in #file:../../instructions/hve-core/rpi-context-discipline.instructions.md for Lean Post-Work Turn behavior, Response Mode Selection, and Subagent Result Handling. + ## Tracking Artifacts All persistent state, session notes, and workflow artifacts are tracked in `.copilot-tracking/` at the root of the workspace when the workflow needs durable records. For simple and most medium requests, the agent may keep research and planning in its own context and skip creating artifact files until task difficulty or workflow needs justify them. diff --git a/.github/agents/hve-core/subagents/implementation-validator.agent.md b/.github/agents/hve-core/subagents/implementation-validator.agent.md index bcff5f27f..1a4954e75 100644 --- a/.github/agents/hve-core/subagents/implementation-validator.agent.md +++ b/.github/agents/hve-core/subagents/implementation-validator.agent.md @@ -199,13 +199,13 @@ During any validation scope, note issues that fall outside predefined categories ## Response Format -Return implementation validation findings including: - -* Relative path to the implementation validation log. -* Validation status: Passed, Partial, Failed, or Blocked. -* Executive summary of findings with severity counts. -* Count of findings by category. -* Top critical and major findings with evidence. -* Scopes completed, skipped, or blocked. -* Suggested areas for additional investigation. -* Clarifying questions requiring parent agent or user input. +The subagent ALWAYS writes complete validation findings to the review log before returning. The chat response is an executive summary only โ€” full fidelity lives on disk. + +Initial chat response โ€” emit at most: +* 1 line: review log file path (the parent re-reads this file when it needs detail). +* 1 line: validation status (Pass / Pass with Warnings / Fail). +* Up to 7 bullet-point findings (each โ‰ค 240 chars). Prioritize blocking issues and regressions. +* Up to 3 clarifying questions, only when blocking. +* 1 short "Full Detail" pointer line: "Re-read `` for complete validation output, test results, and remediation guidance." + +Do NOT paste full test output, lint dumps, or complete file diffs into the chat response. The review log is the source of truth. diff --git a/.github/agents/hve-core/subagents/plan-validator.agent.md b/.github/agents/hve-core/subagents/plan-validator.agent.md index 16422c1fb..517ec5562 100644 --- a/.github/agents/hve-core/subagents/plan-validator.agent.md +++ b/.github/agents/hve-core/subagents/plan-validator.agent.md @@ -94,13 +94,13 @@ When prior planning logs are available, cross-run comparison notes reference res ## Response Format -Return Plan Validation Findings including: - -* Relative path to the Planning Log file. -* Validation status: Validated, Gaps Found, Incomplete, or Blocked. -* Executive summary of identified discrepancies and unaddressed issues with severity and impact. -* Count of DR- items added, updated, or removed in the Planning Log. -* Count of DD- items added, updated, or removed in the Planning Log. -* Coverage summary (internal analysis): counts of covered, partial, missing requirements. -* Severity-ordered discrepancy findings with evidence references from research and plan files. -* Clarifying questions requiring parent agent or user input. +The subagent ALWAYS writes complete validation findings to the Planning Log before returning. The chat response is an executive summary only โ€” full fidelity lives on disk. + +Initial chat response โ€” emit at most: +* 1 line: planning log file path (the parent re-reads this file when it needs detail). +* 1 line: validation status (Pass / Fail โ€” Critical / Fail โ€” Major / Fail โ€” Minor). +* Up to 7 bullet-point severity-ordered findings (each โ‰ค 240 chars). Prioritize critical and major items. +* Up to 3 clarifying questions, only when blocking. +* 1 short "Full Detail" pointer line: "Re-read `` for complete discrepancy details, evidence, and recommended fixes." + +Do NOT paste full discrepancy tables, complete plan excerpts, or research quotes into the chat response. The planning log is the source of truth. diff --git a/.github/agents/hve-core/subagents/researcher-subagent.agent.md b/.github/agents/hve-core/subagents/researcher-subagent.agent.md index b6b6554c0..aad5e0641 100644 --- a/.github/agents/hve-core/subagents/researcher-subagent.agent.md +++ b/.github/agents/hve-core/subagents/researcher-subagent.agent.md @@ -58,10 +58,13 @@ External URLs may still use markdown link syntax. ## Response Format -Return Subagent Research Executive Details and include the following requirements: +The subagent ALWAYS writes complete findings to its subagent file before returning. The chat response is an executive summary only โ€” full fidelity lives on disk. -* The relative path to the subagent research document. -* The status of the subagent research: Complete, In-Progress, Blocked, etc. -* The important details from the subagent research document based on your interpretation. -* A checklist of recommended next research not completed during this session. -* Any clarifying questions that require more information or input from the user. +Initial chat response โ€” emit at most: +* 1 line: subagent file path (the parent re-reads this file when it needs detail). +* 1 line: status (Complete / Blocked / Needs Clarification). +* Up to 7 bullet-point key findings (each โ‰ค 240 chars). Prioritize findings the parent cannot act on without reading the file. +* Up to 3 clarifying questions, only when blocking. +* 1 short "Full Detail" pointer line: "Re-read `` for complete evidence, code blocks, file/line citations, and rejected alternatives." + +Do NOT paste file contents, code blocks, long quotes, or full evidence tables into the chat response. The subagent file is the source of truth. diff --git a/.github/agents/hve-core/subagents/rpi-validator.agent.md b/.github/agents/hve-core/subagents/rpi-validator.agent.md index b77616383..d5c8a0007 100644 --- a/.github/agents/hve-core/subagents/rpi-validator.agent.md +++ b/.github/agents/hve-core/subagents/rpi-validator.agent.md @@ -64,10 +64,13 @@ Create and update the validation document progressively documenting: ## Response Format -Return RPI Validation Executive Details and include the following requirements: +The subagent ALWAYS writes complete validation findings to the review log before returning. The chat response is an executive summary only โ€” full fidelity lives on disk. -* The relative path to the validation document. -* The status of the validation: Passed, Partial, Failed, or Blocked. -* The important details from the validation document based on your interpretation. -* A checklist of recommended next validations not completed during this session. -* Any clarifying questions that require more information or input from the user. +Initial chat response โ€” emit at most: +* 1 line: review log file path (the parent re-reads this file when it needs detail). +* 1 line: validation status (Pass / Fail). +* Up to 7 bullet-point findings (each โ‰ค 240 chars). Prioritize schema violations and missing required sections. +* Up to 3 clarifying questions, only when blocking. +* 1 short "Full Detail" pointer line: "Re-read `` for complete RPI artifact validation details." + +Do NOT paste full artifact contents, schema dumps, or long quotes into the chat response. The review log is the source of truth. diff --git a/.github/agents/hve-core/task-implementor.agent.md b/.github/agents/hve-core/task-implementor.agent.md index 388d76831..ba6af6b20 100644 --- a/.github/agents/hve-core/task-implementor.agent.md +++ b/.github/agents/hve-core/task-implementor.agent.md @@ -56,6 +56,10 @@ The researcher-subagent returns deep research findings: subagent research docume Subagents can run in parallel when investigating independent topics or executing independent phases. +## Context Discipline + +Follow all rules in #file:../../instructions/hve-core/rpi-context-discipline.instructions.md for Lean Post-Work Turn behavior, Response Mode Selection, and Subagent Result Handling. + ## Required Artifacts | Artifact | Path Pattern | Required | diff --git a/.github/agents/hve-core/task-planner.agent.md b/.github/agents/hve-core/task-planner.agent.md index 2de2dd648..2c19356ac 100644 --- a/.github/agents/hve-core/task-planner.agent.md +++ b/.github/agents/hve-core/task-planner.agent.md @@ -58,6 +58,10 @@ Run `Plan Validator` using `runSubagent` or `task`, providing these inputs: Subagents can run in parallel when investigating independent topics or validating independent concerns. +## Context Discipline + +Follow all rules in #file:../../instructions/hve-core/rpi-context-discipline.instructions.md for Lean Post-Work Turn behavior, Response Mode Selection, and Subagent Result Handling. + ## File Locations Planning files reside in `.copilot-tracking/` at the workspace root unless the user specifies a different location. diff --git a/.github/agents/hve-core/task-researcher.agent.md b/.github/agents/hve-core/task-researcher.agent.md index 3739a0338..4783ea3e2 100644 --- a/.github/agents/hve-core/task-researcher.agent.md +++ b/.github/agents/hve-core/task-researcher.agent.md @@ -50,6 +50,10 @@ Run `Researcher Subagent` with `runSubagent` or `task`, and parallelize calls wh Subagents can run in parallel when investigating independent topics or sources. +## Context Discipline + +Follow all rules in #file:../../instructions/hve-core/rpi-context-discipline.instructions.md for Lean Post-Work Turn behavior, Response Mode Selection, and Subagent Result Handling. + ## File Locations Research files reside in `.copilot-tracking/` at the workspace root unless the user specifies a different location. diff --git a/.github/agents/hve-core/task-reviewer.agent.md b/.github/agents/hve-core/task-reviewer.agent.md index b1675147a..8950188a3 100644 --- a/.github/agents/hve-core/task-reviewer.agent.md +++ b/.github/agents/hve-core/task-reviewer.agent.md @@ -37,6 +37,10 @@ Reviews completed implementation work from `.copilot-tracking/` artifacts. Valid * Match `applyTo` patterns from `.github/instructions/` files against changed file types to identify applicable conventions. * Subagents return structured, evidence-based responses with severity levels and can ask clarifying questions rather than guessing. +## Context Discipline + +Follow all rules in #file:../../instructions/hve-core/rpi-context-discipline.instructions.md for Lean Post-Work Turn behavior, Response Mode Selection, and Subagent Result Handling. + ## Review Artifacts | Artifact | Path Pattern | Required | diff --git a/.github/instructions/hve-core/rpi-context-discipline.instructions.md b/.github/instructions/hve-core/rpi-context-discipline.instructions.md new file mode 100644 index 000000000..fff50254b --- /dev/null +++ b/.github/instructions/hve-core/rpi-context-discipline.instructions.md @@ -0,0 +1,31 @@ +--- +description: "Shared context-discipline rules for RPI parent agents โ€” lean post-work turns, response mode selection, and subagent result handling" +applyTo: '**/.copilot-tracking/{research,plans,details,changes,reviews,challenges}/**' +--- + +# RPI Context Discipline + +Shared rules for RPI parent agents (`RPI Agent`, `Task Researcher`, `Task Planner`, `Task Implementor`, `Task Reviewer`, `Task Challenger`) to keep chat context bounded and avoid `/compact` loops. Each parent agent references this file via `#file:` so a single edit propagates. + +## Lean Post-Work Turn + +After any subagent returns, this turn MUST be lean: + +1. Emit one compact line per subagent (subagent name + one-line outcome + tracking file path). +2. Update the relevant `.copilot-tracking/` file via a single edit if needed. +3. Stop. Do NOT re-read large planning, research, or details files in the closing turn. Do NOT re-quote subagent payloads. Do NOT narrate the next phase plan. + +## Response Mode Selection + +Choose the lightest mode that satisfies the request: + +* **Direct** โ€” Answer from this turn's context only. No subagent, no file reads. Use for clarifications, status questions, or queries when the relevant file is already attached. +* **Lightweight** โ€” Single subagent with a focused prompt. Skip re-reading prior phase tracking files. Use for summarizing findings or single-file edits. +* **Standard** โ€” Default behavior: subagent dispatch + tracking-file update + handoff suggestion. +* **Full** โ€” Multiple parallel subagents + cross-phase synthesis. Use only when explicitly requested or when the phase contract requires it. + +## Subagent Result Handling + +* Treat the subagent's chat response as an index, not the full result. +* When a decision (plan structure, phase ordering, accept/reject of an alternative, validation verdict) depends on detail beyond the summary bullets, re-read the subagent file directly and cite specific sections. +* Do not re-read the file gratuitously โ€” only when the next action requires evidence the summary does not contain. diff --git a/collections/hve-core-all.collection.md b/collections/hve-core-all.collection.md index 0403a774a..2edac319d 100644 --- a/collections/hve-core-all.collection.md +++ b/collections/hve-core-all.collection.md @@ -220,6 +220,7 @@ Use this edition when you want access to everything without choosing a focused c | **hve-core/markdown** | Required instructions for creating or editing any Markdown (.md) files | | **hve-core/prompt-builder** | Authoring standards for prompt engineering artifacts including prompts, agents, instructions, and skills | | **hve-core/pull-request** | Required instructions for pull request description generation and optional PR creation using diff analysis, subagent review, and MCP tools | +| **hve-core/rpi-context-discipline** | Shared context-discipline rules for RPI parent agents โ€” lean post-work turns, response mode selection, and subagent result handling | | **hve-core/writing-style** | Required writing style conventions for voice, tone, and language in all markdown content | | **jira/jira-backlog-discovery** | Discovery protocol for Jira backlog management with user-centric, artifact-driven, and JQL-based issue discovery | | **jira/jira-backlog-planning** | Reference specification for Jira backlog management tooling, planning files, search conventions, similarity assessment, and state persistence | diff --git a/collections/hve-core-all.collection.yml b/collections/hve-core-all.collection.yml index f60020053..1c9f57bea 100644 --- a/collections/hve-core-all.collection.yml +++ b/collections/hve-core-all.collection.yml @@ -496,6 +496,8 @@ items: kind: instruction - path: .github/instructions/hve-core/pull-request.instructions.md kind: instruction +- path: .github/instructions/hve-core/rpi-context-discipline.instructions.md + kind: instruction - path: .github/instructions/hve-core/writing-style.instructions.md kind: instruction - path: .github/instructions/jira/jira-backlog-discovery.instructions.md diff --git a/collections/hve-core.collection.md b/collections/hve-core.collection.md index c2fac9a6b..1c5f8421c 100644 --- a/collections/hve-core.collection.md +++ b/collections/hve-core.collection.md @@ -48,15 +48,16 @@ HVE Core provides the flagship RPI (Research, Plan, Implement, Review) workflow ### Instructions -| Name | Description | -|------------------------------|-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------| -| **hve-core/commit-message** | Required instructions for creating all commit messages | -| **hve-core/git-merge** | Required protocol for Git merge, rebase, and rebase --onto workflows with conflict handling and stop controls. | -| **hve-core/markdown** | Required instructions for creating or editing any Markdown (.md) files | -| **hve-core/prompt-builder** | Authoring standards for prompt engineering artifacts including prompts, agents, instructions, and skills | -| **hve-core/pull-request** | Required instructions for pull request description generation and optional PR creation using diff analysis, subagent review, and MCP tools | -| **hve-core/writing-style** | Required writing style conventions for voice, tone, and language in all markdown content | -| **shared/hve-core-location** | Important: hve-core is the repository containing this instruction file; Guidance: if a referenced prompt, instructions, agent, or script is missing in the current directory, fall back to this hve-core location by walking up this file's directory tree. | +| Name | Description | +|-------------------------------------|-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------| +| **hve-core/commit-message** | Required instructions for creating all commit messages | +| **hve-core/git-merge** | Required protocol for Git merge, rebase, and rebase --onto workflows with conflict handling and stop controls. | +| **hve-core/markdown** | Required instructions for creating or editing any Markdown (.md) files | +| **hve-core/prompt-builder** | Authoring standards for prompt engineering artifacts including prompts, agents, instructions, and skills | +| **hve-core/pull-request** | Required instructions for pull request description generation and optional PR creation using diff analysis, subagent review, and MCP tools | +| **hve-core/rpi-context-discipline** | Shared context-discipline rules for RPI parent agents โ€” lean post-work turns, response mode selection, and subagent result handling | +| **hve-core/writing-style** | Required writing style conventions for voice, tone, and language in all markdown content | +| **shared/hve-core-location** | Important: hve-core is the repository containing this instruction file; Guidance: if a referenced prompt, instructions, agent, or script is missing in the current directory, fall back to this hve-core location by walking up this file's directory tree. | ### Skills diff --git a/collections/hve-core.collection.yml b/collections/hve-core.collection.yml index 879f3a210..690d3dd3b 100644 --- a/collections/hve-core.collection.yml +++ b/collections/hve-core.collection.yml @@ -103,6 +103,8 @@ items: kind: instruction - path: .github/instructions/hve-core/pull-request.instructions.md kind: instruction + - path: .github/instructions/hve-core/rpi-context-discipline.instructions.md + kind: instruction - path: .github/instructions/shared/hve-core-location.instructions.md kind: instruction # Skills diff --git a/collections/project-planning.collection.md b/collections/project-planning.collection.md index 292f67dba..e0943fc9d 100644 --- a/collections/project-planning.collection.md +++ b/collections/project-planning.collection.md @@ -46,6 +46,7 @@ Create architecture decision records, requirements documents, and diagrams โ€” a | Name | Description | |------------------------------------------|-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------| +| **hve-core/rpi-context-discipline** | Shared context-discipline rules for RPI parent agents โ€” lean post-work turns, response mode selection, and subagent result handling | | **rai-planning/rai-backlog-handoff** | RAI review and backlog handoff for Phase 6: review rubric, RAI review summary, dual-format backlog generation | | **rai-planning/rai-capture-coaching** | Exploration-first questioning techniques for RAI capture mode adapted from Design Thinking research methods | | **rai-planning/rai-identity** | RAI Planner identity, 6-phase orchestration, state management, and session recovery | diff --git a/collections/project-planning.collection.yml b/collections/project-planning.collection.yml index e2ad4d2e8..748d8d2b5 100644 --- a/collections/project-planning.collection.yml +++ b/collections/project-planning.collection.yml @@ -92,6 +92,8 @@ items: kind: prompt maturity: experimental # Instructions + - path: .github/instructions/hve-core/rpi-context-discipline.instructions.md + kind: instruction - path: .github/instructions/rai-planning/rai-backlog-handoff.instructions.md kind: instruction maturity: experimental diff --git a/plugins/hve-core-all/README.md b/plugins/hve-core-all/README.md index edab24775..f16c55b2f 100644 --- a/plugins/hve-core-all/README.md +++ b/plugins/hve-core-all/README.md @@ -498,6 +498,7 @@ copilot plugin install hve-core-all@hve-core | markdown.instructions | Required instructions for creating or editing any Markdown (.md) files | | prompt-builder.instructions | Authoring standards for prompt engineering artifacts including prompts, agents, instructions, and skills | | pull-request.instructions | Required instructions for pull request description generation and optional PR creation using diff analysis, subagent review, and MCP tools - Brought to you by microsoft/hve-core | +| rpi-context-discipline.instructions | Shared context-discipline rules for RPI parent agents โ€” lean post-work turns, response mode selection, and subagent result handling | | writing-style.instructions | Required writing style conventions for voice, tone, and language in all markdown content | | jira-backlog-discovery.instructions | Discovery protocol for Jira backlog management with user-centric, artifact-driven, and JQL-based issue discovery - Brought to you by microsoft/hve-core | | jira-backlog-planning.instructions | Reference specification for Jira backlog management tooling, planning files, search conventions, similarity assessment, and state persistence - Brought to you by microsoft/hve-core | diff --git a/plugins/hve-core-all/instructions/hve-core/rpi-context-discipline.instructions.md b/plugins/hve-core-all/instructions/hve-core/rpi-context-discipline.instructions.md new file mode 120000 index 000000000..ac4b81e95 --- /dev/null +++ b/plugins/hve-core-all/instructions/hve-core/rpi-context-discipline.instructions.md @@ -0,0 +1 @@ +../../../../.github/instructions/hve-core/rpi-context-discipline.instructions.md \ No newline at end of file diff --git a/plugins/hve-core/README.md b/plugins/hve-core/README.md index a5168e197..f1b174639 100644 --- a/plugins/hve-core/README.md +++ b/plugins/hve-core/README.md @@ -125,15 +125,16 @@ copilot plugin install hve-core@hve-core ## Instructions -| Instruction | Description | -|--------------------------------|-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------| -| writing-style.instructions | Required writing style conventions for voice, tone, and language in all markdown content | -| markdown.instructions | Required instructions for creating or editing any Markdown (.md) files | -| commit-message.instructions | Required instructions for creating all commit messages - Brought to you by microsoft/hve-core | -| prompt-builder.instructions | Authoring standards for prompt engineering artifacts including prompts, agents, instructions, and skills | -| git-merge.instructions | Required protocol for Git merge, rebase, and rebase --onto workflows with conflict handling and stop controls. | -| pull-request.instructions | Required instructions for pull request description generation and optional PR creation using diff analysis, subagent review, and MCP tools - Brought to you by microsoft/hve-core | -| hve-core-location.instructions | Important: hve-core is the repository containing this instruction file; Guidance: if a referenced prompt, instructions, agent, or script is missing in the current directory, fall back to this hve-core location by walking up this file's directory tree. | +| Instruction | Description | +|-------------------------------------|-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------| +| writing-style.instructions | Required writing style conventions for voice, tone, and language in all markdown content | +| markdown.instructions | Required instructions for creating or editing any Markdown (.md) files | +| commit-message.instructions | Required instructions for creating all commit messages - Brought to you by microsoft/hve-core | +| prompt-builder.instructions | Authoring standards for prompt engineering artifacts including prompts, agents, instructions, and skills | +| git-merge.instructions | Required protocol for Git merge, rebase, and rebase --onto workflows with conflict handling and stop controls. | +| pull-request.instructions | Required instructions for pull request description generation and optional PR creation using diff analysis, subagent review, and MCP tools - Brought to you by microsoft/hve-core | +| rpi-context-discipline.instructions | Shared context-discipline rules for RPI parent agents โ€” lean post-work turns, response mode selection, and subagent result handling | +| hve-core-location.instructions | Important: hve-core is the repository containing this instruction file; Guidance: if a referenced prompt, instructions, agent, or script is missing in the current directory, fall back to this hve-core location by walking up this file's directory tree. | ## Skills diff --git a/plugins/hve-core/instructions/hve-core/rpi-context-discipline.instructions.md b/plugins/hve-core/instructions/hve-core/rpi-context-discipline.instructions.md new file mode 120000 index 000000000..ac4b81e95 --- /dev/null +++ b/plugins/hve-core/instructions/hve-core/rpi-context-discipline.instructions.md @@ -0,0 +1 @@ +../../../../.github/instructions/hve-core/rpi-context-discipline.instructions.md \ No newline at end of file diff --git a/plugins/project-planning/README.md b/plugins/project-planning/README.md index 0da6f2bc9..3a65dcf8c 100644 --- a/plugins/project-planning/README.md +++ b/plugins/project-planning/README.md @@ -127,6 +127,7 @@ copilot plugin install project-planning@hve-core | Instruction | Description | |--------------------------------------|-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------| +| rpi-context-discipline.instructions | Shared context-discipline rules for RPI parent agents โ€” lean post-work turns, response mode selection, and subagent result handling | | rai-backlog-handoff.instructions | RAI review and backlog handoff for Phase 6: review rubric, RAI review summary, dual-format backlog generation | | rai-identity.instructions | RAI Planner identity, 6-phase orchestration, state management, and session recovery - Brought to you by microsoft/hve-core | | rai-impact-assessment.instructions | RAI impact assessment for Phase 5: control surface taxonomy, evidence register, tradeoff documentation, and work item generation - Brought to you by microsoft/hve-core | diff --git a/plugins/project-planning/instructions/hve-core/rpi-context-discipline.instructions.md b/plugins/project-planning/instructions/hve-core/rpi-context-discipline.instructions.md new file mode 120000 index 000000000..ac4b81e95 --- /dev/null +++ b/plugins/project-planning/instructions/hve-core/rpi-context-discipline.instructions.md @@ -0,0 +1 @@ +../../../../.github/instructions/hve-core/rpi-context-discipline.instructions.md \ No newline at end of file From 64691681266d4cf79c292d87bf11a747f7515965 Mon Sep 17 00:00:00 2001 From: katriendg Date: Thu, 30 Apr 2026 13:44:16 +0000 Subject: [PATCH 2/9] feat(agents): enhance plan-validator and researcher subagents with additional output details MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit - add count of DR- and DD- items in plan-validator - include relative path to research document in researcher subagent - add checklist of recommended next steps for both subagents ๐Ÿ” - Generated by Copilot --- .github/agents/hve-core/subagents/plan-validator.agent.md | 2 ++ .github/agents/hve-core/subagents/researcher-subagent.agent.md | 2 ++ .github/agents/hve-core/subagents/rpi-validator.agent.md | 2 ++ 3 files changed, 6 insertions(+) diff --git a/.github/agents/hve-core/subagents/plan-validator.agent.md b/.github/agents/hve-core/subagents/plan-validator.agent.md index 517ec5562..e9a00a2f2 100644 --- a/.github/agents/hve-core/subagents/plan-validator.agent.md +++ b/.github/agents/hve-core/subagents/plan-validator.agent.md @@ -99,6 +99,8 @@ The subagent ALWAYS writes complete validation findings to the Planning Log befo Initial chat response โ€” emit at most: * 1 line: planning log file path (the parent re-reads this file when it needs detail). * 1 line: validation status (Pass / Fail โ€” Critical / Fail โ€” Major / Fail โ€” Minor). +* Count of DR- items added, updated, or removed in the Planning Log. +* Count of DD- items added, updated, or removed in the Planning Log. * Up to 7 bullet-point severity-ordered findings (each โ‰ค 240 chars). Prioritize critical and major items. * Up to 3 clarifying questions, only when blocking. * 1 short "Full Detail" pointer line: "Re-read `` for complete discrepancy details, evidence, and recommended fixes." diff --git a/.github/agents/hve-core/subagents/researcher-subagent.agent.md b/.github/agents/hve-core/subagents/researcher-subagent.agent.md index aad5e0641..174efd1fc 100644 --- a/.github/agents/hve-core/subagents/researcher-subagent.agent.md +++ b/.github/agents/hve-core/subagents/researcher-subagent.agent.md @@ -61,7 +61,9 @@ External URLs may still use markdown link syntax. The subagent ALWAYS writes complete findings to its subagent file before returning. The chat response is an executive summary only โ€” full fidelity lives on disk. Initial chat response โ€” emit at most: +* The relative path to the subagent research document. * 1 line: subagent file path (the parent re-reads this file when it needs detail). +* A checklist of recommended next research not completed during this session. * 1 line: status (Complete / Blocked / Needs Clarification). * Up to 7 bullet-point key findings (each โ‰ค 240 chars). Prioritize findings the parent cannot act on without reading the file. * Up to 3 clarifying questions, only when blocking. diff --git a/.github/agents/hve-core/subagents/rpi-validator.agent.md b/.github/agents/hve-core/subagents/rpi-validator.agent.md index d5c8a0007..250d70b93 100644 --- a/.github/agents/hve-core/subagents/rpi-validator.agent.md +++ b/.github/agents/hve-core/subagents/rpi-validator.agent.md @@ -67,6 +67,8 @@ Create and update the validation document progressively documenting: The subagent ALWAYS writes complete validation findings to the review log before returning. The chat response is an executive summary only โ€” full fidelity lives on disk. Initial chat response โ€” emit at most: +* The relative path to the subagent research document. +* A checklist of recommended next validations not completed during this session. * 1 line: review log file path (the parent re-reads this file when it needs detail). * 1 line: validation status (Pass / Fail). * Up to 7 bullet-point findings (each โ‰ค 240 chars). Prioritize schema violations and missing required sections. From b63642e61800dcfa716089783eda34d02f7973d1 Mon Sep 17 00:00:00 2001 From: katriendg Date: Thu, 30 Apr 2026 13:47:32 +0000 Subject: [PATCH 3/9] feat(instructions): add shared context-discipline rules for RPI parent agents MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit ๐Ÿ”’ - Generated by Copilot --- plugins/hve-core-all/README.md | 1 + plugins/hve-core/README.md | 19 ++++++++++--------- plugins/project-planning/README.md | 1 + 3 files changed, 12 insertions(+), 9 deletions(-) diff --git a/plugins/hve-core-all/README.md b/plugins/hve-core-all/README.md index f16c55b2f..af7352b17 100644 --- a/plugins/hve-core-all/README.md +++ b/plugins/hve-core-all/README.md @@ -227,6 +227,7 @@ Use this edition when you want access to everything without choosing a focused c | **hve-core/markdown** | Required instructions for creating or editing any Markdown (.md) files | | **hve-core/prompt-builder** | Authoring standards for prompt engineering artifacts including prompts, agents, instructions, and skills | | **hve-core/pull-request** | Required instructions for pull request description generation and optional PR creation using diff analysis, subagent review, and MCP tools | +| **hve-core/rpi-context-discipline** | Shared context-discipline rules for RPI parent agents โ€” lean post-work turns, response mode selection, and subagent result handling | | **hve-core/writing-style** | Required writing style conventions for voice, tone, and language in all markdown content | | **jira/jira-backlog-discovery** | Discovery protocol for Jira backlog management with user-centric, artifact-driven, and JQL-based issue discovery | | **jira/jira-backlog-planning** | Reference specification for Jira backlog management tooling, planning files, search conventions, similarity assessment, and state persistence | diff --git a/plugins/hve-core/README.md b/plugins/hve-core/README.md index f1b174639..01c6a5ae6 100644 --- a/plugins/hve-core/README.md +++ b/plugins/hve-core/README.md @@ -55,15 +55,16 @@ HVE Core provides the flagship RPI (Research, Plan, Implement, Review) workflow ### Instructions -| Name | Description | -|------------------------------|-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------| -| **hve-core/commit-message** | Required instructions for creating all commit messages | -| **hve-core/git-merge** | Required protocol for Git merge, rebase, and rebase --onto workflows with conflict handling and stop controls. | -| **hve-core/markdown** | Required instructions for creating or editing any Markdown (.md) files | -| **hve-core/prompt-builder** | Authoring standards for prompt engineering artifacts including prompts, agents, instructions, and skills | -| **hve-core/pull-request** | Required instructions for pull request description generation and optional PR creation using diff analysis, subagent review, and MCP tools | -| **hve-core/writing-style** | Required writing style conventions for voice, tone, and language in all markdown content | -| **shared/hve-core-location** | Important: hve-core is the repository containing this instruction file; Guidance: if a referenced prompt, instructions, agent, or script is missing in the current directory, fall back to this hve-core location by walking up this file's directory tree. | +| Name | Description | +|-------------------------------------|-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------| +| **hve-core/commit-message** | Required instructions for creating all commit messages | +| **hve-core/git-merge** | Required protocol for Git merge, rebase, and rebase --onto workflows with conflict handling and stop controls. | +| **hve-core/markdown** | Required instructions for creating or editing any Markdown (.md) files | +| **hve-core/prompt-builder** | Authoring standards for prompt engineering artifacts including prompts, agents, instructions, and skills | +| **hve-core/pull-request** | Required instructions for pull request description generation and optional PR creation using diff analysis, subagent review, and MCP tools | +| **hve-core/rpi-context-discipline** | Shared context-discipline rules for RPI parent agents โ€” lean post-work turns, response mode selection, and subagent result handling | +| **hve-core/writing-style** | Required writing style conventions for voice, tone, and language in all markdown content | +| **shared/hve-core-location** | Important: hve-core is the repository containing this instruction file; Guidance: if a referenced prompt, instructions, agent, or script is missing in the current directory, fall back to this hve-core location by walking up this file's directory tree. | ### Skills diff --git a/plugins/project-planning/README.md b/plugins/project-planning/README.md index 3a65dcf8c..e76bae756 100644 --- a/plugins/project-planning/README.md +++ b/plugins/project-planning/README.md @@ -53,6 +53,7 @@ Create architecture decision records, requirements documents, and diagrams โ€” a | Name | Description | |------------------------------------------|-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------| +| **hve-core/rpi-context-discipline** | Shared context-discipline rules for RPI parent agents โ€” lean post-work turns, response mode selection, and subagent result handling | | **rai-planning/rai-backlog-handoff** | RAI review and backlog handoff for Phase 6: review rubric, RAI review summary, dual-format backlog generation | | **rai-planning/rai-capture-coaching** | Exploration-first questioning techniques for RAI capture mode adapted from Design Thinking research methods | | **rai-planning/rai-identity** | RAI Planner identity, 6-phase orchestration, state management, and session recovery | From 34f489860252a1aba329b82934927801509e895e Mon Sep 17 00:00:00 2001 From: katriendg Date: Mon, 4 May 2026 09:25:43 +0000 Subject: [PATCH 4/9] refactor(agents): tighten RPI context-discipline description and align subagent response formats MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit - shorten rpi-context-discipline description to fit 120-char target - drop duplicated path bullet from researcher and rpi-validator response formats - reorder plan-validator response to put findings before DR-/DD- delta line - regenerate hve-core, hve-core-all, and project-planning plugin READMEs โœ๏ธ - Generated by Copilot --- .github/agents/hve-core/subagents/plan-validator.agent.md | 3 +-- .../agents/hve-core/subagents/researcher-subagent.agent.md | 3 +-- .github/agents/hve-core/subagents/rpi-validator.agent.md | 3 +-- .../hve-core/rpi-context-discipline.instructions.md | 2 +- collections/hve-core-all.collection.md | 2 +- collections/hve-core.collection.md | 2 +- collections/project-planning.collection.md | 2 +- plugins/hve-core-all/README.md | 4 ++-- plugins/hve-core/README.md | 4 ++-- plugins/project-planning/README.md | 4 ++-- 10 files changed, 13 insertions(+), 16 deletions(-) diff --git a/.github/agents/hve-core/subagents/plan-validator.agent.md b/.github/agents/hve-core/subagents/plan-validator.agent.md index e9a00a2f2..9daf09482 100644 --- a/.github/agents/hve-core/subagents/plan-validator.agent.md +++ b/.github/agents/hve-core/subagents/plan-validator.agent.md @@ -99,9 +99,8 @@ The subagent ALWAYS writes complete validation findings to the Planning Log befo Initial chat response โ€” emit at most: * 1 line: planning log file path (the parent re-reads this file when it needs detail). * 1 line: validation status (Pass / Fail โ€” Critical / Fail โ€” Major / Fail โ€” Minor). -* Count of DR- items added, updated, or removed in the Planning Log. -* Count of DD- items added, updated, or removed in the Planning Log. * Up to 7 bullet-point severity-ordered findings (each โ‰ค 240 chars). Prioritize critical and major items. +* 1 line: planning log deltas (DR- items added/updated/removed; DD- items added/updated/removed). * Up to 3 clarifying questions, only when blocking. * 1 short "Full Detail" pointer line: "Re-read `` for complete discrepancy details, evidence, and recommended fixes." diff --git a/.github/agents/hve-core/subagents/researcher-subagent.agent.md b/.github/agents/hve-core/subagents/researcher-subagent.agent.md index 174efd1fc..0356b6746 100644 --- a/.github/agents/hve-core/subagents/researcher-subagent.agent.md +++ b/.github/agents/hve-core/subagents/researcher-subagent.agent.md @@ -61,11 +61,10 @@ External URLs may still use markdown link syntax. The subagent ALWAYS writes complete findings to its subagent file before returning. The chat response is an executive summary only โ€” full fidelity lives on disk. Initial chat response โ€” emit at most: -* The relative path to the subagent research document. * 1 line: subagent file path (the parent re-reads this file when it needs detail). -* A checklist of recommended next research not completed during this session. * 1 line: status (Complete / Blocked / Needs Clarification). * Up to 7 bullet-point key findings (each โ‰ค 240 chars). Prioritize findings the parent cannot act on without reading the file. +* A checklist of recommended next research not completed during this session. * Up to 3 clarifying questions, only when blocking. * 1 short "Full Detail" pointer line: "Re-read `` for complete evidence, code blocks, file/line citations, and rejected alternatives." diff --git a/.github/agents/hve-core/subagents/rpi-validator.agent.md b/.github/agents/hve-core/subagents/rpi-validator.agent.md index 250d70b93..a8ab3d8fe 100644 --- a/.github/agents/hve-core/subagents/rpi-validator.agent.md +++ b/.github/agents/hve-core/subagents/rpi-validator.agent.md @@ -67,11 +67,10 @@ Create and update the validation document progressively documenting: The subagent ALWAYS writes complete validation findings to the review log before returning. The chat response is an executive summary only โ€” full fidelity lives on disk. Initial chat response โ€” emit at most: -* The relative path to the subagent research document. -* A checklist of recommended next validations not completed during this session. * 1 line: review log file path (the parent re-reads this file when it needs detail). * 1 line: validation status (Pass / Fail). * Up to 7 bullet-point findings (each โ‰ค 240 chars). Prioritize schema violations and missing required sections. +* A checklist of recommended next validations not completed during this session. * Up to 3 clarifying questions, only when blocking. * 1 short "Full Detail" pointer line: "Re-read `` for complete RPI artifact validation details." diff --git a/.github/instructions/hve-core/rpi-context-discipline.instructions.md b/.github/instructions/hve-core/rpi-context-discipline.instructions.md index fff50254b..3500aa4cb 100644 --- a/.github/instructions/hve-core/rpi-context-discipline.instructions.md +++ b/.github/instructions/hve-core/rpi-context-discipline.instructions.md @@ -1,5 +1,5 @@ --- -description: "Shared context-discipline rules for RPI parent agents โ€” lean post-work turns, response mode selection, and subagent result handling" +description: "Context-discipline rules for RPI parent agents: lean post-work turns, response modes, subagent results" applyTo: '**/.copilot-tracking/{research,plans,details,changes,reviews,challenges}/**' --- diff --git a/collections/hve-core-all.collection.md b/collections/hve-core-all.collection.md index f7abbc266..ce4e928a6 100644 --- a/collections/hve-core-all.collection.md +++ b/collections/hve-core-all.collection.md @@ -220,7 +220,7 @@ Use this edition when you want access to everything without choosing a focused c | **hve-core/markdown** | Required instructions for creating or editing any Markdown (.md) files | | **hve-core/prompt-builder** | Authoring standards for prompt engineering artifacts including prompts, agents, instructions, and skills | | **hve-core/pull-request** | Required instructions for pull request description generation and optional PR creation using diff analysis, subagent review, and MCP tools | -| **hve-core/rpi-context-discipline** | Shared context-discipline rules for RPI parent agents โ€” lean post-work turns, response mode selection, and subagent result handling | +| **hve-core/rpi-context-discipline** | Context-discipline rules for RPI parent agents: lean post-work turns, response modes, subagent results | | **hve-core/writing-style** | Required writing style conventions for voice, tone, and language in all markdown content | | **jira/jira-backlog-discovery** | Discovery protocol for Jira backlog management with user-centric, artifact-driven, and JQL-based issue discovery | | **jira/jira-backlog-planning** | Reference specification for Jira backlog management tooling, planning files, search conventions, similarity assessment, and state persistence | diff --git a/collections/hve-core.collection.md b/collections/hve-core.collection.md index 1c5f8421c..c3a0a6bb4 100644 --- a/collections/hve-core.collection.md +++ b/collections/hve-core.collection.md @@ -55,7 +55,7 @@ HVE Core provides the flagship RPI (Research, Plan, Implement, Review) workflow | **hve-core/markdown** | Required instructions for creating or editing any Markdown (.md) files | | **hve-core/prompt-builder** | Authoring standards for prompt engineering artifacts including prompts, agents, instructions, and skills | | **hve-core/pull-request** | Required instructions for pull request description generation and optional PR creation using diff analysis, subagent review, and MCP tools | -| **hve-core/rpi-context-discipline** | Shared context-discipline rules for RPI parent agents โ€” lean post-work turns, response mode selection, and subagent result handling | +| **hve-core/rpi-context-discipline** | Context-discipline rules for RPI parent agents: lean post-work turns, response modes, subagent results | | **hve-core/writing-style** | Required writing style conventions for voice, tone, and language in all markdown content | | **shared/hve-core-location** | Important: hve-core is the repository containing this instruction file; Guidance: if a referenced prompt, instructions, agent, or script is missing in the current directory, fall back to this hve-core location by walking up this file's directory tree. | diff --git a/collections/project-planning.collection.md b/collections/project-planning.collection.md index e0943fc9d..06c072076 100644 --- a/collections/project-planning.collection.md +++ b/collections/project-planning.collection.md @@ -46,7 +46,7 @@ Create architecture decision records, requirements documents, and diagrams โ€” a | Name | Description | |------------------------------------------|-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------| -| **hve-core/rpi-context-discipline** | Shared context-discipline rules for RPI parent agents โ€” lean post-work turns, response mode selection, and subagent result handling | +| **hve-core/rpi-context-discipline** | Context-discipline rules for RPI parent agents: lean post-work turns, response modes, subagent results | | **rai-planning/rai-backlog-handoff** | RAI review and backlog handoff for Phase 6: review rubric, RAI review summary, dual-format backlog generation | | **rai-planning/rai-capture-coaching** | Exploration-first questioning techniques for RAI capture mode adapted from Design Thinking research methods | | **rai-planning/rai-identity** | RAI Planner identity, 6-phase orchestration, state management, and session recovery | diff --git a/plugins/hve-core-all/README.md b/plugins/hve-core-all/README.md index 0b8eda96a..791dd7952 100644 --- a/plugins/hve-core-all/README.md +++ b/plugins/hve-core-all/README.md @@ -227,7 +227,7 @@ Use this edition when you want access to everything without choosing a focused c | **hve-core/markdown** | Required instructions for creating or editing any Markdown (.md) files | | **hve-core/prompt-builder** | Authoring standards for prompt engineering artifacts including prompts, agents, instructions, and skills | | **hve-core/pull-request** | Required instructions for pull request description generation and optional PR creation using diff analysis, subagent review, and MCP tools | -| **hve-core/rpi-context-discipline** | Shared context-discipline rules for RPI parent agents โ€” lean post-work turns, response mode selection, and subagent result handling | +| **hve-core/rpi-context-discipline** | Context-discipline rules for RPI parent agents: lean post-work turns, response modes, subagent results | | **hve-core/writing-style** | Required writing style conventions for voice, tone, and language in all markdown content | | **jira/jira-backlog-discovery** | Discovery protocol for Jira backlog management with user-centric, artifact-driven, and JQL-based issue discovery | | **jira/jira-backlog-planning** | Reference specification for Jira backlog management tooling, planning files, search conventions, similarity assessment, and state persistence | @@ -500,7 +500,7 @@ copilot plugin install hve-core-all@hve-core | markdown.instructions | Required instructions for creating or editing any Markdown (.md) files | | prompt-builder.instructions | Authoring standards for prompt engineering artifacts including prompts, agents, instructions, and skills | | pull-request.instructions | Required instructions for pull request description generation and optional PR creation using diff analysis, subagent review, and MCP tools - Brought to you by microsoft/hve-core | -| rpi-context-discipline.instructions | Shared context-discipline rules for RPI parent agents โ€” lean post-work turns, response mode selection, and subagent result handling | +| rpi-context-discipline.instructions | Context-discipline rules for RPI parent agents: lean post-work turns, response modes, subagent results | | writing-style.instructions | Required writing style conventions for voice, tone, and language in all markdown content | | jira-backlog-discovery.instructions | Discovery protocol for Jira backlog management with user-centric, artifact-driven, and JQL-based issue discovery - Brought to you by microsoft/hve-core | | jira-backlog-planning.instructions | Reference specification for Jira backlog management tooling, planning files, search conventions, similarity assessment, and state persistence - Brought to you by microsoft/hve-core | diff --git a/plugins/hve-core/README.md b/plugins/hve-core/README.md index 01c6a5ae6..b824fdef9 100644 --- a/plugins/hve-core/README.md +++ b/plugins/hve-core/README.md @@ -62,7 +62,7 @@ HVE Core provides the flagship RPI (Research, Plan, Implement, Review) workflow | **hve-core/markdown** | Required instructions for creating or editing any Markdown (.md) files | | **hve-core/prompt-builder** | Authoring standards for prompt engineering artifacts including prompts, agents, instructions, and skills | | **hve-core/pull-request** | Required instructions for pull request description generation and optional PR creation using diff analysis, subagent review, and MCP tools | -| **hve-core/rpi-context-discipline** | Shared context-discipline rules for RPI parent agents โ€” lean post-work turns, response mode selection, and subagent result handling | +| **hve-core/rpi-context-discipline** | Context-discipline rules for RPI parent agents: lean post-work turns, response modes, subagent results | | **hve-core/writing-style** | Required writing style conventions for voice, tone, and language in all markdown content | | **shared/hve-core-location** | Important: hve-core is the repository containing this instruction file; Guidance: if a referenced prompt, instructions, agent, or script is missing in the current directory, fall back to this hve-core location by walking up this file's directory tree. | @@ -134,7 +134,7 @@ copilot plugin install hve-core@hve-core | prompt-builder.instructions | Authoring standards for prompt engineering artifacts including prompts, agents, instructions, and skills | | git-merge.instructions | Required protocol for Git merge, rebase, and rebase --onto workflows with conflict handling and stop controls. | | pull-request.instructions | Required instructions for pull request description generation and optional PR creation using diff analysis, subagent review, and MCP tools - Brought to you by microsoft/hve-core | -| rpi-context-discipline.instructions | Shared context-discipline rules for RPI parent agents โ€” lean post-work turns, response mode selection, and subagent result handling | +| rpi-context-discipline.instructions | Context-discipline rules for RPI parent agents: lean post-work turns, response modes, subagent results | | hve-core-location.instructions | Important: hve-core is the repository containing this instruction file; Guidance: if a referenced prompt, instructions, agent, or script is missing in the current directory, fall back to this hve-core location by walking up this file's directory tree. | ## Skills diff --git a/plugins/project-planning/README.md b/plugins/project-planning/README.md index e76bae756..856688c43 100644 --- a/plugins/project-planning/README.md +++ b/plugins/project-planning/README.md @@ -53,7 +53,7 @@ Create architecture decision records, requirements documents, and diagrams โ€” a | Name | Description | |------------------------------------------|-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------| -| **hve-core/rpi-context-discipline** | Shared context-discipline rules for RPI parent agents โ€” lean post-work turns, response mode selection, and subagent result handling | +| **hve-core/rpi-context-discipline** | Context-discipline rules for RPI parent agents: lean post-work turns, response modes, subagent results | | **rai-planning/rai-backlog-handoff** | RAI review and backlog handoff for Phase 6: review rubric, RAI review summary, dual-format backlog generation | | **rai-planning/rai-capture-coaching** | Exploration-first questioning techniques for RAI capture mode adapted from Design Thinking research methods | | **rai-planning/rai-identity** | RAI Planner identity, 6-phase orchestration, state management, and session recovery | @@ -128,7 +128,7 @@ copilot plugin install project-planning@hve-core | Instruction | Description | |--------------------------------------|-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------| -| rpi-context-discipline.instructions | Shared context-discipline rules for RPI parent agents โ€” lean post-work turns, response mode selection, and subagent result handling | +| rpi-context-discipline.instructions | Context-discipline rules for RPI parent agents: lean post-work turns, response modes, subagent results | | rai-backlog-handoff.instructions | RAI review and backlog handoff for Phase 6: review rubric, RAI review summary, dual-format backlog generation | | rai-identity.instructions | RAI Planner identity, 6-phase orchestration, state management, and session recovery - Brought to you by microsoft/hve-core | | rai-impact-assessment.instructions | RAI impact assessment for Phase 5: control surface taxonomy, evidence register, tradeoff documentation, and work item generation - Brought to you by microsoft/hve-core | From 023b41cd68cbdeccd5702d6ffe7a2059f0f052ce Mon Sep 17 00:00:00 2001 From: katriendg Date: Mon, 4 May 2026 09:34:01 +0000 Subject: [PATCH 5/9] refactor(instructions): remove Task Challenger from RPI parent agents description and update applyTo path MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit ๐Ÿ”ง - Generated by Copilot Co-authored-by: Copilot --- .../hve-core/rpi-context-discipline.instructions.md | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) diff --git a/.github/instructions/hve-core/rpi-context-discipline.instructions.md b/.github/instructions/hve-core/rpi-context-discipline.instructions.md index 3500aa4cb..a06447491 100644 --- a/.github/instructions/hve-core/rpi-context-discipline.instructions.md +++ b/.github/instructions/hve-core/rpi-context-discipline.instructions.md @@ -1,11 +1,13 @@ --- description: "Context-discipline rules for RPI parent agents: lean post-work turns, response modes, subagent results" -applyTo: '**/.copilot-tracking/{research,plans,details,changes,reviews,challenges}/**' +applyTo: '**/.copilot-tracking/{research,plans,details,changes,reviews}/**' --- # RPI Context Discipline -Shared rules for RPI parent agents (`RPI Agent`, `Task Researcher`, `Task Planner`, `Task Implementor`, `Task Reviewer`, `Task Challenger`) to keep chat context bounded and avoid `/compact` loops. Each parent agent references this file via `#file:` so a single edit propagates. +Shared rules for RPI parent agents that dispatch subagents (`RPI Agent`, `Task Researcher`, `Task Planner`, `Task Implementor`, `Task Reviewer`) to keep chat context bounded and avoid `/compact` loops. Each parent agent references this file via `#file:` so a single edit propagates. + +`Task Challenger` is intentionally out of scope: it does not dispatch subagents and already enforces a stricter one-question-per-response protocol. ## Lean Post-Work Turn From b3d55bda26b213fe7bf8b5d6cedd089f36ef2318 Mon Sep 17 00:00:00 2001 From: katriendg Date: Mon, 4 May 2026 12:38:04 +0000 Subject: [PATCH 6/9] chore: address PR comments Co-authored-by: Copilot --- .../rpi-context-discipline.instructions.md | 14 ++++++++------ plugins/hve-core-all/README.md | 2 +- plugins/hve-core/README.md | 2 +- plugins/project-planning/README.md | 2 +- 4 files changed, 11 insertions(+), 9 deletions(-) diff --git a/.github/instructions/hve-core/rpi-context-discipline.instructions.md b/.github/instructions/hve-core/rpi-context-discipline.instructions.md index a06447491..eb02790b9 100644 --- a/.github/instructions/hve-core/rpi-context-discipline.instructions.md +++ b/.github/instructions/hve-core/rpi-context-discipline.instructions.md @@ -1,5 +1,5 @@ --- -description: "Context-discipline rules for RPI parent agents: lean post-work turns, response modes, subagent results" +description: "Context-discipline rules for RPI parent agents: lean post-work turns, response modes, subagent results - Brought to you by microsoft/hve-core" applyTo: '**/.copilot-tracking/{research,plans,details,changes,reviews}/**' --- @@ -21,13 +21,15 @@ After any subagent returns, this turn MUST be lean: Choose the lightest mode that satisfies the request: -* **Direct** โ€” Answer from this turn's context only. No subagent, no file reads. Use for clarifications, status questions, or queries when the relevant file is already attached. -* **Lightweight** โ€” Single subagent with a focused prompt. Skip re-reading prior phase tracking files. Use for summarizing findings or single-file edits. -* **Standard** โ€” Default behavior: subagent dispatch + tracking-file update + handoff suggestion. -* **Full** โ€” Multiple parallel subagents + cross-phase synthesis. Use only when explicitly requested or when the phase contract requires it. +| Mode | When to use | +|-------------|--------------------------------------------------------------------------------------------------------------------------------------------------------------------| +| Direct | Answer from this turn's context only. No subagent, no file reads. Use for clarifications, status questions, or queries when the relevant file is already attached. | +| Lightweight | Single subagent with a focused prompt. Skip re-reading prior phase tracking files. Use for summarizing findings or single-file edits. | +| Standard | Default behavior: subagent dispatch, tracking-file update, and handoff suggestion. | +| Full | Multiple parallel subagents and cross-phase synthesis. Use only when explicitly requested or when the phase contract requires it. | ## Subagent Result Handling * Treat the subagent's chat response as an index, not the full result. * When a decision (plan structure, phase ordering, accept/reject of an alternative, validation verdict) depends on detail beyond the summary bullets, re-read the subagent file directly and cite specific sections. -* Do not re-read the file gratuitously โ€” only when the next action requires evidence the summary does not contain. +* Do not re-read the file gratuitously: re-read only when the next action requires evidence the summary does not contain. diff --git a/plugins/hve-core-all/README.md b/plugins/hve-core-all/README.md index 791dd7952..e0306c63b 100644 --- a/plugins/hve-core-all/README.md +++ b/plugins/hve-core-all/README.md @@ -500,7 +500,7 @@ copilot plugin install hve-core-all@hve-core | markdown.instructions | Required instructions for creating or editing any Markdown (.md) files | | prompt-builder.instructions | Authoring standards for prompt engineering artifacts including prompts, agents, instructions, and skills | | pull-request.instructions | Required instructions for pull request description generation and optional PR creation using diff analysis, subagent review, and MCP tools - Brought to you by microsoft/hve-core | -| rpi-context-discipline.instructions | Context-discipline rules for RPI parent agents: lean post-work turns, response modes, subagent results | +| rpi-context-discipline.instructions | Context-discipline rules for RPI parent agents: lean post-work turns, response modes, subagent results - Brought to you by microsoft/hve-core | | writing-style.instructions | Required writing style conventions for voice, tone, and language in all markdown content | | jira-backlog-discovery.instructions | Discovery protocol for Jira backlog management with user-centric, artifact-driven, and JQL-based issue discovery - Brought to you by microsoft/hve-core | | jira-backlog-planning.instructions | Reference specification for Jira backlog management tooling, planning files, search conventions, similarity assessment, and state persistence - Brought to you by microsoft/hve-core | diff --git a/plugins/hve-core/README.md b/plugins/hve-core/README.md index b824fdef9..e7218681c 100644 --- a/plugins/hve-core/README.md +++ b/plugins/hve-core/README.md @@ -134,7 +134,7 @@ copilot plugin install hve-core@hve-core | prompt-builder.instructions | Authoring standards for prompt engineering artifacts including prompts, agents, instructions, and skills | | git-merge.instructions | Required protocol for Git merge, rebase, and rebase --onto workflows with conflict handling and stop controls. | | pull-request.instructions | Required instructions for pull request description generation and optional PR creation using diff analysis, subagent review, and MCP tools - Brought to you by microsoft/hve-core | -| rpi-context-discipline.instructions | Context-discipline rules for RPI parent agents: lean post-work turns, response modes, subagent results | +| rpi-context-discipline.instructions | Context-discipline rules for RPI parent agents: lean post-work turns, response modes, subagent results - Brought to you by microsoft/hve-core | | hve-core-location.instructions | Important: hve-core is the repository containing this instruction file; Guidance: if a referenced prompt, instructions, agent, or script is missing in the current directory, fall back to this hve-core location by walking up this file's directory tree. | ## Skills diff --git a/plugins/project-planning/README.md b/plugins/project-planning/README.md index 856688c43..e0775a01e 100644 --- a/plugins/project-planning/README.md +++ b/plugins/project-planning/README.md @@ -128,7 +128,7 @@ copilot plugin install project-planning@hve-core | Instruction | Description | |--------------------------------------|-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------| -| rpi-context-discipline.instructions | Context-discipline rules for RPI parent agents: lean post-work turns, response modes, subagent results | +| rpi-context-discipline.instructions | Context-discipline rules for RPI parent agents: lean post-work turns, response modes, subagent results - Brought to you by microsoft/hve-core | | rai-backlog-handoff.instructions | RAI review and backlog handoff for Phase 6: review rubric, RAI review summary, dual-format backlog generation | | rai-identity.instructions | RAI Planner identity, 6-phase orchestration, state management, and session recovery - Brought to you by microsoft/hve-core | | rai-impact-assessment.instructions | RAI impact assessment for Phase 5: control surface taxonomy, evidence register, tradeoff documentation, and work item generation - Brought to you by microsoft/hve-core | From 47ab0564150e63cf90beaf9d6d44d12b37889430 Mon Sep 17 00:00:00 2001 From: katriendg Date: Tue, 5 May 2026 06:53:49 +0000 Subject: [PATCH 7/9] refactor(agents): update context discipline rules for RPI agents and remove separate instructions file MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit - streamline response mode selection and subagent result handling - clarify lean post-work turn requirements - remove deprecated context-discipline instructions from various collections ๐Ÿ”ง - Generated by Copilot --- .github/agents/hve-core/rpi-agent.agent.md | 21 +++++++++- .../implementation-validator.agent.md | 4 +- .../subagents/plan-validator.agent.md | 6 +-- .../subagents/researcher-subagent.agent.md | 4 +- .../hve-core/subagents/rpi-validator.agent.md | 4 +- .../agents/hve-core/task-implementor.agent.md | 21 +++++++++- .github/agents/hve-core/task-planner.agent.md | 21 +++++++++- .../agents/hve-core/task-researcher.agent.md | 21 +++++++++- .../agents/hve-core/task-reviewer.agent.md | 21 +++++++++- .../rpi-context-discipline.instructions.md | 35 ----------------- collections/hve-core-all.collection.md | 1 - collections/hve-core-all.collection.yml | 2 - collections/hve-core.collection.md | 19 +++++----- collections/hve-core.collection.yml | 2 - collections/project-planning.collection.md | 1 - collections/project-planning.collection.yml | 2 - plugins/hve-core-all/README.md | 2 - .../rpi-context-discipline.instructions.md | 1 - plugins/hve-core/README.md | 38 +++++++++---------- .../rpi-context-discipline.instructions.md | 1 - plugins/project-planning/README.md | 2 - .../rpi-context-discipline.instructions.md | 1 - 22 files changed, 136 insertions(+), 94 deletions(-) delete mode 100644 .github/instructions/hve-core/rpi-context-discipline.instructions.md delete mode 120000 plugins/hve-core-all/instructions/hve-core/rpi-context-discipline.instructions.md delete mode 120000 plugins/hve-core/instructions/hve-core/rpi-context-discipline.instructions.md delete mode 120000 plugins/project-planning/instructions/hve-core/rpi-context-discipline.instructions.md diff --git a/.github/agents/hve-core/rpi-agent.agent.md b/.github/agents/hve-core/rpi-agent.agent.md index 49b2d588d..25f03ad5e 100644 --- a/.github/agents/hve-core/rpi-agent.agent.md +++ b/.github/agents/hve-core/rpi-agent.agent.md @@ -93,7 +93,26 @@ Treat the phase guidance below as operating defaults rather than ceremony. Deleg ## Context Discipline -Follow all rules in #file:../../instructions/hve-core/rpi-context-discipline.instructions.md for Lean Post-Work Turn behavior, Response Mode Selection, and Subagent Result Handling. +After any subagent returns, this turn MUST be lean: + +1. Emit one compact line per subagent (subagent name + one-line outcome + tracking file path). +2. Update the relevant `.copilot-tracking/` file via a single edit if needed. +3. Stop. Do NOT re-read large planning, research, or details files in the closing turn. Do NOT re-quote subagent payloads. Do NOT narrate the next phase plan. + +Choose the lightest response mode that satisfies the request: + +| Mode | When to use | +|-------------|--------------------------------------------------------------------------------------------------------------------------------------------------------------------| +| Direct | Answer from this turn's context only. No subagent, no file reads. Use for clarifications, status questions, or queries when the relevant file is already attached. | +| Lightweight | Single subagent with a focused prompt. Skip re-reading prior phase tracking files. Use for summarizing findings or single-file edits. | +| Standard | Default behavior: subagent dispatch, tracking-file update, and handoff suggestion. | +| Full | Multiple parallel subagents and cross-phase synthesis. Use only when explicitly requested or when the phase contract requires it. | + +Subagent result handling: + +* Treat the subagent's chat response as an index, not the full result. +* When a decision (plan structure, phase ordering, accept/reject of an alternative, validation verdict) depends on detail beyond the summary bullets, re-read the subagent file directly and cite specific sections. +* Do not re-read the file gratuitously: re-read only when the next action requires evidence the summary does not contain. ## Tracking Artifacts diff --git a/.github/agents/hve-core/subagents/implementation-validator.agent.md b/.github/agents/hve-core/subagents/implementation-validator.agent.md index 1a4954e75..2198701ca 100644 --- a/.github/agents/hve-core/subagents/implementation-validator.agent.md +++ b/.github/agents/hve-core/subagents/implementation-validator.agent.md @@ -199,9 +199,9 @@ During any validation scope, note issues that fall outside predefined categories ## Response Format -The subagent ALWAYS writes complete validation findings to the review log before returning. The chat response is an executive summary only โ€” full fidelity lives on disk. +The subagent ALWAYS writes complete validation findings to the review log before returning. The chat response is an executive summary only. Full fidelity lives on disk. -Initial chat response โ€” emit at most: +Initial chat response, emit at most: * 1 line: review log file path (the parent re-reads this file when it needs detail). * 1 line: validation status (Pass / Pass with Warnings / Fail). * Up to 7 bullet-point findings (each โ‰ค 240 chars). Prioritize blocking issues and regressions. diff --git a/.github/agents/hve-core/subagents/plan-validator.agent.md b/.github/agents/hve-core/subagents/plan-validator.agent.md index 9daf09482..52a943971 100644 --- a/.github/agents/hve-core/subagents/plan-validator.agent.md +++ b/.github/agents/hve-core/subagents/plan-validator.agent.md @@ -94,11 +94,11 @@ When prior planning logs are available, cross-run comparison notes reference res ## Response Format -The subagent ALWAYS writes complete validation findings to the Planning Log before returning. The chat response is an executive summary only โ€” full fidelity lives on disk. +The subagent ALWAYS writes complete validation findings to the Planning Log before returning. The chat response is an executive summary only. Full fidelity lives on disk. -Initial chat response โ€” emit at most: +Initial chat response, emit at most: * 1 line: planning log file path (the parent re-reads this file when it needs detail). -* 1 line: validation status (Pass / Fail โ€” Critical / Fail โ€” Major / Fail โ€” Minor). +* 1 line: validation status (Pass / Fail - Critical / Fail - Major / Fail - Minor). * Up to 7 bullet-point severity-ordered findings (each โ‰ค 240 chars). Prioritize critical and major items. * 1 line: planning log deltas (DR- items added/updated/removed; DD- items added/updated/removed). * Up to 3 clarifying questions, only when blocking. diff --git a/.github/agents/hve-core/subagents/researcher-subagent.agent.md b/.github/agents/hve-core/subagents/researcher-subagent.agent.md index 0356b6746..9b3046d97 100644 --- a/.github/agents/hve-core/subagents/researcher-subagent.agent.md +++ b/.github/agents/hve-core/subagents/researcher-subagent.agent.md @@ -58,9 +58,9 @@ External URLs may still use markdown link syntax. ## Response Format -The subagent ALWAYS writes complete findings to its subagent file before returning. The chat response is an executive summary only โ€” full fidelity lives on disk. +The subagent ALWAYS writes complete findings to its subagent file before returning. The chat response is an executive summary only. Full fidelity lives on disk. -Initial chat response โ€” emit at most: +Initial chat response, emit at most: * 1 line: subagent file path (the parent re-reads this file when it needs detail). * 1 line: status (Complete / Blocked / Needs Clarification). * Up to 7 bullet-point key findings (each โ‰ค 240 chars). Prioritize findings the parent cannot act on without reading the file. diff --git a/.github/agents/hve-core/subagents/rpi-validator.agent.md b/.github/agents/hve-core/subagents/rpi-validator.agent.md index a8ab3d8fe..fbc01caf0 100644 --- a/.github/agents/hve-core/subagents/rpi-validator.agent.md +++ b/.github/agents/hve-core/subagents/rpi-validator.agent.md @@ -64,9 +64,9 @@ Create and update the validation document progressively documenting: ## Response Format -The subagent ALWAYS writes complete validation findings to the review log before returning. The chat response is an executive summary only โ€” full fidelity lives on disk. +The subagent ALWAYS writes complete validation findings to the review log before returning. The chat response is an executive summary only. Full fidelity lives on disk. -Initial chat response โ€” emit at most: +Initial chat response, emit at most: * 1 line: review log file path (the parent re-reads this file when it needs detail). * 1 line: validation status (Pass / Fail). * Up to 7 bullet-point findings (each โ‰ค 240 chars). Prioritize schema violations and missing required sections. diff --git a/.github/agents/hve-core/task-implementor.agent.md b/.github/agents/hve-core/task-implementor.agent.md index ba6af6b20..76623b47c 100644 --- a/.github/agents/hve-core/task-implementor.agent.md +++ b/.github/agents/hve-core/task-implementor.agent.md @@ -58,7 +58,26 @@ Subagents can run in parallel when investigating independent topics or executing ## Context Discipline -Follow all rules in #file:../../instructions/hve-core/rpi-context-discipline.instructions.md for Lean Post-Work Turn behavior, Response Mode Selection, and Subagent Result Handling. +After any subagent returns, this turn MUST be lean: + +1. Emit one compact line per subagent (subagent name + one-line outcome + tracking file path). +2. Update the relevant `.copilot-tracking/` file via a single edit if needed. +3. Stop. Do NOT re-read large planning, research, or details files in the closing turn. Do NOT re-quote subagent payloads. Do NOT narrate the next phase plan. + +Choose the lightest response mode that satisfies the request: + +| Mode | When to use | +|-------------|--------------------------------------------------------------------------------------------------------------------------------------------------------------------| +| Direct | Answer from this turn's context only. No subagent, no file reads. Use for clarifications, status questions, or queries when the relevant file is already attached. | +| Lightweight | Single subagent with a focused prompt. Skip re-reading prior phase tracking files. Use for summarizing findings or single-file edits. | +| Standard | Default behavior: subagent dispatch, tracking-file update, and handoff suggestion. | +| Full | Multiple parallel subagents and cross-phase synthesis. Use only when explicitly requested or when the phase contract requires it. | + +Subagent result handling: + +* Treat the subagent's chat response as an index, not the full result. +* When a decision (plan structure, phase ordering, accept/reject of an alternative, validation verdict) depends on detail beyond the summary bullets, re-read the subagent file directly and cite specific sections. +* Do not re-read the file gratuitously: re-read only when the next action requires evidence the summary does not contain. ## Required Artifacts diff --git a/.github/agents/hve-core/task-planner.agent.md b/.github/agents/hve-core/task-planner.agent.md index 2c19356ac..ddd773366 100644 --- a/.github/agents/hve-core/task-planner.agent.md +++ b/.github/agents/hve-core/task-planner.agent.md @@ -60,7 +60,26 @@ Subagents can run in parallel when investigating independent topics or validatin ## Context Discipline -Follow all rules in #file:../../instructions/hve-core/rpi-context-discipline.instructions.md for Lean Post-Work Turn behavior, Response Mode Selection, and Subagent Result Handling. +After any subagent returns, this turn MUST be lean: + +1. Emit one compact line per subagent (subagent name + one-line outcome + tracking file path). +2. Update the relevant `.copilot-tracking/` file via a single edit if needed. +3. Stop. Do NOT re-read large planning, research, or details files in the closing turn. Do NOT re-quote subagent payloads. Do NOT narrate the next phase plan. + +Choose the lightest response mode that satisfies the request: + +| Mode | When to use | +|-------------|--------------------------------------------------------------------------------------------------------------------------------------------------------------------| +| Direct | Answer from this turn's context only. No subagent, no file reads. Use for clarifications, status questions, or queries when the relevant file is already attached. | +| Lightweight | Single subagent with a focused prompt. Skip re-reading prior phase tracking files. Use for summarizing findings or single-file edits. | +| Standard | Default behavior: subagent dispatch, tracking-file update, and handoff suggestion. | +| Full | Multiple parallel subagents and cross-phase synthesis. Use only when explicitly requested or when the phase contract requires it. | + +Subagent result handling: + +* Treat the subagent's chat response as an index, not the full result. +* When a decision (plan structure, phase ordering, accept/reject of an alternative, validation verdict) depends on detail beyond the summary bullets, re-read the subagent file directly and cite specific sections. +* Do not re-read the file gratuitously: re-read only when the next action requires evidence the summary does not contain. ## File Locations diff --git a/.github/agents/hve-core/task-researcher.agent.md b/.github/agents/hve-core/task-researcher.agent.md index 4783ea3e2..759fa674e 100644 --- a/.github/agents/hve-core/task-researcher.agent.md +++ b/.github/agents/hve-core/task-researcher.agent.md @@ -52,7 +52,26 @@ Subagents can run in parallel when investigating independent topics or sources. ## Context Discipline -Follow all rules in #file:../../instructions/hve-core/rpi-context-discipline.instructions.md for Lean Post-Work Turn behavior, Response Mode Selection, and Subagent Result Handling. +After any subagent returns, this turn MUST be lean: + +1. Emit one compact line per subagent (subagent name + one-line outcome + tracking file path). +2. Update the relevant `.copilot-tracking/` file via a single edit if needed. +3. Stop. Do NOT re-read large planning, research, or details files in the closing turn. Do NOT re-quote subagent payloads. Do NOT narrate the next phase plan. + +Choose the lightest response mode that satisfies the request: + +| Mode | When to use | +|-------------|--------------------------------------------------------------------------------------------------------------------------------------------------------------------| +| Direct | Answer from this turn's context only. No subagent, no file reads. Use for clarifications, status questions, or queries when the relevant file is already attached. | +| Lightweight | Single subagent with a focused prompt. Skip re-reading prior phase tracking files. Use for summarizing findings or single-file edits. | +| Standard | Default behavior: subagent dispatch, tracking-file update, and handoff suggestion. | +| Full | Multiple parallel subagents and cross-phase synthesis. Use only when explicitly requested or when the phase contract requires it. | + +Subagent result handling: + +* Treat the subagent's chat response as an index, not the full result. +* When a decision (plan structure, phase ordering, accept/reject of an alternative, validation verdict) depends on detail beyond the summary bullets, re-read the subagent file directly and cite specific sections. +* Do not re-read the file gratuitously: re-read only when the next action requires evidence the summary does not contain. ## File Locations diff --git a/.github/agents/hve-core/task-reviewer.agent.md b/.github/agents/hve-core/task-reviewer.agent.md index 8950188a3..2202040e5 100644 --- a/.github/agents/hve-core/task-reviewer.agent.md +++ b/.github/agents/hve-core/task-reviewer.agent.md @@ -39,7 +39,26 @@ Reviews completed implementation work from `.copilot-tracking/` artifacts. Valid ## Context Discipline -Follow all rules in #file:../../instructions/hve-core/rpi-context-discipline.instructions.md for Lean Post-Work Turn behavior, Response Mode Selection, and Subagent Result Handling. +After any subagent returns, this turn MUST be lean: + +1. Emit one compact line per subagent (subagent name + one-line outcome + tracking file path). +2. Update the relevant `.copilot-tracking/` file via a single edit if needed. +3. Stop. Do NOT re-read large planning, research, or details files in the closing turn. Do NOT re-quote subagent payloads. Do NOT narrate the next phase plan. + +Choose the lightest response mode that satisfies the request: + +| Mode | When to use | +|-------------|--------------------------------------------------------------------------------------------------------------------------------------------------------------------| +| Direct | Answer from this turn's context only. No subagent, no file reads. Use for clarifications, status questions, or queries when the relevant file is already attached. | +| Lightweight | Single subagent with a focused prompt. Skip re-reading prior phase tracking files. Use for summarizing findings or single-file edits. | +| Standard | Default behavior: subagent dispatch, tracking-file update, and handoff suggestion. | +| Full | Multiple parallel subagents and cross-phase synthesis. Use only when explicitly requested or when the phase contract requires it. | + +Subagent result handling: + +* Treat the subagent's chat response as an index, not the full result. +* When a decision (plan structure, phase ordering, accept/reject of an alternative, validation verdict) depends on detail beyond the summary bullets, re-read the subagent file directly and cite specific sections. +* Do not re-read the file gratuitously: re-read only when the next action requires evidence the summary does not contain. ## Review Artifacts diff --git a/.github/instructions/hve-core/rpi-context-discipline.instructions.md b/.github/instructions/hve-core/rpi-context-discipline.instructions.md deleted file mode 100644 index eb02790b9..000000000 --- a/.github/instructions/hve-core/rpi-context-discipline.instructions.md +++ /dev/null @@ -1,35 +0,0 @@ ---- -description: "Context-discipline rules for RPI parent agents: lean post-work turns, response modes, subagent results - Brought to you by microsoft/hve-core" -applyTo: '**/.copilot-tracking/{research,plans,details,changes,reviews}/**' ---- - -# RPI Context Discipline - -Shared rules for RPI parent agents that dispatch subagents (`RPI Agent`, `Task Researcher`, `Task Planner`, `Task Implementor`, `Task Reviewer`) to keep chat context bounded and avoid `/compact` loops. Each parent agent references this file via `#file:` so a single edit propagates. - -`Task Challenger` is intentionally out of scope: it does not dispatch subagents and already enforces a stricter one-question-per-response protocol. - -## Lean Post-Work Turn - -After any subagent returns, this turn MUST be lean: - -1. Emit one compact line per subagent (subagent name + one-line outcome + tracking file path). -2. Update the relevant `.copilot-tracking/` file via a single edit if needed. -3. Stop. Do NOT re-read large planning, research, or details files in the closing turn. Do NOT re-quote subagent payloads. Do NOT narrate the next phase plan. - -## Response Mode Selection - -Choose the lightest mode that satisfies the request: - -| Mode | When to use | -|-------------|--------------------------------------------------------------------------------------------------------------------------------------------------------------------| -| Direct | Answer from this turn's context only. No subagent, no file reads. Use for clarifications, status questions, or queries when the relevant file is already attached. | -| Lightweight | Single subagent with a focused prompt. Skip re-reading prior phase tracking files. Use for summarizing findings or single-file edits. | -| Standard | Default behavior: subagent dispatch, tracking-file update, and handoff suggestion. | -| Full | Multiple parallel subagents and cross-phase synthesis. Use only when explicitly requested or when the phase contract requires it. | - -## Subagent Result Handling - -* Treat the subagent's chat response as an index, not the full result. -* When a decision (plan structure, phase ordering, accept/reject of an alternative, validation verdict) depends on detail beyond the summary bullets, re-read the subagent file directly and cite specific sections. -* Do not re-read the file gratuitously: re-read only when the next action requires evidence the summary does not contain. diff --git a/collections/hve-core-all.collection.md b/collections/hve-core-all.collection.md index ce4e928a6..ce848da12 100644 --- a/collections/hve-core-all.collection.md +++ b/collections/hve-core-all.collection.md @@ -220,7 +220,6 @@ Use this edition when you want access to everything without choosing a focused c | **hve-core/markdown** | Required instructions for creating or editing any Markdown (.md) files | | **hve-core/prompt-builder** | Authoring standards for prompt engineering artifacts including prompts, agents, instructions, and skills | | **hve-core/pull-request** | Required instructions for pull request description generation and optional PR creation using diff analysis, subagent review, and MCP tools | -| **hve-core/rpi-context-discipline** | Context-discipline rules for RPI parent agents: lean post-work turns, response modes, subagent results | | **hve-core/writing-style** | Required writing style conventions for voice, tone, and language in all markdown content | | **jira/jira-backlog-discovery** | Discovery protocol for Jira backlog management with user-centric, artifact-driven, and JQL-based issue discovery | | **jira/jira-backlog-planning** | Reference specification for Jira backlog management tooling, planning files, search conventions, similarity assessment, and state persistence | diff --git a/collections/hve-core-all.collection.yml b/collections/hve-core-all.collection.yml index b06d0bb78..eb93015a8 100644 --- a/collections/hve-core-all.collection.yml +++ b/collections/hve-core-all.collection.yml @@ -496,8 +496,6 @@ items: kind: instruction - path: .github/instructions/hve-core/pull-request.instructions.md kind: instruction -- path: .github/instructions/hve-core/rpi-context-discipline.instructions.md - kind: instruction - path: .github/instructions/hve-core/writing-style.instructions.md kind: instruction - path: .github/instructions/jira/jira-backlog-discovery.instructions.md diff --git a/collections/hve-core.collection.md b/collections/hve-core.collection.md index c3a0a6bb4..c2fac9a6b 100644 --- a/collections/hve-core.collection.md +++ b/collections/hve-core.collection.md @@ -48,16 +48,15 @@ HVE Core provides the flagship RPI (Research, Plan, Implement, Review) workflow ### Instructions -| Name | Description | -|-------------------------------------|-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------| -| **hve-core/commit-message** | Required instructions for creating all commit messages | -| **hve-core/git-merge** | Required protocol for Git merge, rebase, and rebase --onto workflows with conflict handling and stop controls. | -| **hve-core/markdown** | Required instructions for creating or editing any Markdown (.md) files | -| **hve-core/prompt-builder** | Authoring standards for prompt engineering artifacts including prompts, agents, instructions, and skills | -| **hve-core/pull-request** | Required instructions for pull request description generation and optional PR creation using diff analysis, subagent review, and MCP tools | -| **hve-core/rpi-context-discipline** | Context-discipline rules for RPI parent agents: lean post-work turns, response modes, subagent results | -| **hve-core/writing-style** | Required writing style conventions for voice, tone, and language in all markdown content | -| **shared/hve-core-location** | Important: hve-core is the repository containing this instruction file; Guidance: if a referenced prompt, instructions, agent, or script is missing in the current directory, fall back to this hve-core location by walking up this file's directory tree. | +| Name | Description | +|------------------------------|-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------| +| **hve-core/commit-message** | Required instructions for creating all commit messages | +| **hve-core/git-merge** | Required protocol for Git merge, rebase, and rebase --onto workflows with conflict handling and stop controls. | +| **hve-core/markdown** | Required instructions for creating or editing any Markdown (.md) files | +| **hve-core/prompt-builder** | Authoring standards for prompt engineering artifacts including prompts, agents, instructions, and skills | +| **hve-core/pull-request** | Required instructions for pull request description generation and optional PR creation using diff analysis, subagent review, and MCP tools | +| **hve-core/writing-style** | Required writing style conventions for voice, tone, and language in all markdown content | +| **shared/hve-core-location** | Important: hve-core is the repository containing this instruction file; Guidance: if a referenced prompt, instructions, agent, or script is missing in the current directory, fall back to this hve-core location by walking up this file's directory tree. | ### Skills diff --git a/collections/hve-core.collection.yml b/collections/hve-core.collection.yml index 690d3dd3b..879f3a210 100644 --- a/collections/hve-core.collection.yml +++ b/collections/hve-core.collection.yml @@ -103,8 +103,6 @@ items: kind: instruction - path: .github/instructions/hve-core/pull-request.instructions.md kind: instruction - - path: .github/instructions/hve-core/rpi-context-discipline.instructions.md - kind: instruction - path: .github/instructions/shared/hve-core-location.instructions.md kind: instruction # Skills diff --git a/collections/project-planning.collection.md b/collections/project-planning.collection.md index 06c072076..292f67dba 100644 --- a/collections/project-planning.collection.md +++ b/collections/project-planning.collection.md @@ -46,7 +46,6 @@ Create architecture decision records, requirements documents, and diagrams โ€” a | Name | Description | |------------------------------------------|-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------| -| **hve-core/rpi-context-discipline** | Context-discipline rules for RPI parent agents: lean post-work turns, response modes, subagent results | | **rai-planning/rai-backlog-handoff** | RAI review and backlog handoff for Phase 6: review rubric, RAI review summary, dual-format backlog generation | | **rai-planning/rai-capture-coaching** | Exploration-first questioning techniques for RAI capture mode adapted from Design Thinking research methods | | **rai-planning/rai-identity** | RAI Planner identity, 6-phase orchestration, state management, and session recovery | diff --git a/collections/project-planning.collection.yml b/collections/project-planning.collection.yml index 748d8d2b5..e2ad4d2e8 100644 --- a/collections/project-planning.collection.yml +++ b/collections/project-planning.collection.yml @@ -92,8 +92,6 @@ items: kind: prompt maturity: experimental # Instructions - - path: .github/instructions/hve-core/rpi-context-discipline.instructions.md - kind: instruction - path: .github/instructions/rai-planning/rai-backlog-handoff.instructions.md kind: instruction maturity: experimental diff --git a/plugins/hve-core-all/README.md b/plugins/hve-core-all/README.md index e0306c63b..c11f5efd7 100644 --- a/plugins/hve-core-all/README.md +++ b/plugins/hve-core-all/README.md @@ -227,7 +227,6 @@ Use this edition when you want access to everything without choosing a focused c | **hve-core/markdown** | Required instructions for creating or editing any Markdown (.md) files | | **hve-core/prompt-builder** | Authoring standards for prompt engineering artifacts including prompts, agents, instructions, and skills | | **hve-core/pull-request** | Required instructions for pull request description generation and optional PR creation using diff analysis, subagent review, and MCP tools | -| **hve-core/rpi-context-discipline** | Context-discipline rules for RPI parent agents: lean post-work turns, response modes, subagent results | | **hve-core/writing-style** | Required writing style conventions for voice, tone, and language in all markdown content | | **jira/jira-backlog-discovery** | Discovery protocol for Jira backlog management with user-centric, artifact-driven, and JQL-based issue discovery | | **jira/jira-backlog-planning** | Reference specification for Jira backlog management tooling, planning files, search conventions, similarity assessment, and state persistence | @@ -500,7 +499,6 @@ copilot plugin install hve-core-all@hve-core | markdown.instructions | Required instructions for creating or editing any Markdown (.md) files | | prompt-builder.instructions | Authoring standards for prompt engineering artifacts including prompts, agents, instructions, and skills | | pull-request.instructions | Required instructions for pull request description generation and optional PR creation using diff analysis, subagent review, and MCP tools - Brought to you by microsoft/hve-core | -| rpi-context-discipline.instructions | Context-discipline rules for RPI parent agents: lean post-work turns, response modes, subagent results - Brought to you by microsoft/hve-core | | writing-style.instructions | Required writing style conventions for voice, tone, and language in all markdown content | | jira-backlog-discovery.instructions | Discovery protocol for Jira backlog management with user-centric, artifact-driven, and JQL-based issue discovery - Brought to you by microsoft/hve-core | | jira-backlog-planning.instructions | Reference specification for Jira backlog management tooling, planning files, search conventions, similarity assessment, and state persistence - Brought to you by microsoft/hve-core | diff --git a/plugins/hve-core-all/instructions/hve-core/rpi-context-discipline.instructions.md b/plugins/hve-core-all/instructions/hve-core/rpi-context-discipline.instructions.md deleted file mode 120000 index ac4b81e95..000000000 --- a/plugins/hve-core-all/instructions/hve-core/rpi-context-discipline.instructions.md +++ /dev/null @@ -1 +0,0 @@ -../../../../.github/instructions/hve-core/rpi-context-discipline.instructions.md \ No newline at end of file diff --git a/plugins/hve-core/README.md b/plugins/hve-core/README.md index e7218681c..a5168e197 100644 --- a/plugins/hve-core/README.md +++ b/plugins/hve-core/README.md @@ -55,16 +55,15 @@ HVE Core provides the flagship RPI (Research, Plan, Implement, Review) workflow ### Instructions -| Name | Description | -|-------------------------------------|-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------| -| **hve-core/commit-message** | Required instructions for creating all commit messages | -| **hve-core/git-merge** | Required protocol for Git merge, rebase, and rebase --onto workflows with conflict handling and stop controls. | -| **hve-core/markdown** | Required instructions for creating or editing any Markdown (.md) files | -| **hve-core/prompt-builder** | Authoring standards for prompt engineering artifacts including prompts, agents, instructions, and skills | -| **hve-core/pull-request** | Required instructions for pull request description generation and optional PR creation using diff analysis, subagent review, and MCP tools | -| **hve-core/rpi-context-discipline** | Context-discipline rules for RPI parent agents: lean post-work turns, response modes, subagent results | -| **hve-core/writing-style** | Required writing style conventions for voice, tone, and language in all markdown content | -| **shared/hve-core-location** | Important: hve-core is the repository containing this instruction file; Guidance: if a referenced prompt, instructions, agent, or script is missing in the current directory, fall back to this hve-core location by walking up this file's directory tree. | +| Name | Description | +|------------------------------|-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------| +| **hve-core/commit-message** | Required instructions for creating all commit messages | +| **hve-core/git-merge** | Required protocol for Git merge, rebase, and rebase --onto workflows with conflict handling and stop controls. | +| **hve-core/markdown** | Required instructions for creating or editing any Markdown (.md) files | +| **hve-core/prompt-builder** | Authoring standards for prompt engineering artifacts including prompts, agents, instructions, and skills | +| **hve-core/pull-request** | Required instructions for pull request description generation and optional PR creation using diff analysis, subagent review, and MCP tools | +| **hve-core/writing-style** | Required writing style conventions for voice, tone, and language in all markdown content | +| **shared/hve-core-location** | Important: hve-core is the repository containing this instruction file; Guidance: if a referenced prompt, instructions, agent, or script is missing in the current directory, fall back to this hve-core location by walking up this file's directory tree. | ### Skills @@ -126,16 +125,15 @@ copilot plugin install hve-core@hve-core ## Instructions -| Instruction | Description | -|-------------------------------------|-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------| -| writing-style.instructions | Required writing style conventions for voice, tone, and language in all markdown content | -| markdown.instructions | Required instructions for creating or editing any Markdown (.md) files | -| commit-message.instructions | Required instructions for creating all commit messages - Brought to you by microsoft/hve-core | -| prompt-builder.instructions | Authoring standards for prompt engineering artifacts including prompts, agents, instructions, and skills | -| git-merge.instructions | Required protocol for Git merge, rebase, and rebase --onto workflows with conflict handling and stop controls. | -| pull-request.instructions | Required instructions for pull request description generation and optional PR creation using diff analysis, subagent review, and MCP tools - Brought to you by microsoft/hve-core | -| rpi-context-discipline.instructions | Context-discipline rules for RPI parent agents: lean post-work turns, response modes, subagent results - Brought to you by microsoft/hve-core | -| hve-core-location.instructions | Important: hve-core is the repository containing this instruction file; Guidance: if a referenced prompt, instructions, agent, or script is missing in the current directory, fall back to this hve-core location by walking up this file's directory tree. | +| Instruction | Description | +|--------------------------------|-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------| +| writing-style.instructions | Required writing style conventions for voice, tone, and language in all markdown content | +| markdown.instructions | Required instructions for creating or editing any Markdown (.md) files | +| commit-message.instructions | Required instructions for creating all commit messages - Brought to you by microsoft/hve-core | +| prompt-builder.instructions | Authoring standards for prompt engineering artifacts including prompts, agents, instructions, and skills | +| git-merge.instructions | Required protocol for Git merge, rebase, and rebase --onto workflows with conflict handling and stop controls. | +| pull-request.instructions | Required instructions for pull request description generation and optional PR creation using diff analysis, subagent review, and MCP tools - Brought to you by microsoft/hve-core | +| hve-core-location.instructions | Important: hve-core is the repository containing this instruction file; Guidance: if a referenced prompt, instructions, agent, or script is missing in the current directory, fall back to this hve-core location by walking up this file's directory tree. | ## Skills diff --git a/plugins/hve-core/instructions/hve-core/rpi-context-discipline.instructions.md b/plugins/hve-core/instructions/hve-core/rpi-context-discipline.instructions.md deleted file mode 120000 index ac4b81e95..000000000 --- a/plugins/hve-core/instructions/hve-core/rpi-context-discipline.instructions.md +++ /dev/null @@ -1 +0,0 @@ -../../../../.github/instructions/hve-core/rpi-context-discipline.instructions.md \ No newline at end of file diff --git a/plugins/project-planning/README.md b/plugins/project-planning/README.md index e0775a01e..0da6f2bc9 100644 --- a/plugins/project-planning/README.md +++ b/plugins/project-planning/README.md @@ -53,7 +53,6 @@ Create architecture decision records, requirements documents, and diagrams โ€” a | Name | Description | |------------------------------------------|-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------| -| **hve-core/rpi-context-discipline** | Context-discipline rules for RPI parent agents: lean post-work turns, response modes, subagent results | | **rai-planning/rai-backlog-handoff** | RAI review and backlog handoff for Phase 6: review rubric, RAI review summary, dual-format backlog generation | | **rai-planning/rai-capture-coaching** | Exploration-first questioning techniques for RAI capture mode adapted from Design Thinking research methods | | **rai-planning/rai-identity** | RAI Planner identity, 6-phase orchestration, state management, and session recovery | @@ -128,7 +127,6 @@ copilot plugin install project-planning@hve-core | Instruction | Description | |--------------------------------------|-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------| -| rpi-context-discipline.instructions | Context-discipline rules for RPI parent agents: lean post-work turns, response modes, subagent results - Brought to you by microsoft/hve-core | | rai-backlog-handoff.instructions | RAI review and backlog handoff for Phase 6: review rubric, RAI review summary, dual-format backlog generation | | rai-identity.instructions | RAI Planner identity, 6-phase orchestration, state management, and session recovery - Brought to you by microsoft/hve-core | | rai-impact-assessment.instructions | RAI impact assessment for Phase 5: control surface taxonomy, evidence register, tradeoff documentation, and work item generation - Brought to you by microsoft/hve-core | diff --git a/plugins/project-planning/instructions/hve-core/rpi-context-discipline.instructions.md b/plugins/project-planning/instructions/hve-core/rpi-context-discipline.instructions.md deleted file mode 120000 index ac4b81e95..000000000 --- a/plugins/project-planning/instructions/hve-core/rpi-context-discipline.instructions.md +++ /dev/null @@ -1 +0,0 @@ -../../../../.github/instructions/hve-core/rpi-context-discipline.instructions.md \ No newline at end of file From 536714900b81dd8afcdd3742ac1cd054ce629a18 Mon Sep 17 00:00:00 2001 From: katriendg Date: Tue, 5 May 2026 07:10:10 +0000 Subject: [PATCH 8/9] refactor(agents): limit checklist items to 5 for researcher and RPI validator subagents MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit ๐Ÿ”ง - Generated by Copilot --- .github/agents/hve-core/subagents/researcher-subagent.agent.md | 2 +- .github/agents/hve-core/subagents/rpi-validator.agent.md | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) diff --git a/.github/agents/hve-core/subagents/researcher-subagent.agent.md b/.github/agents/hve-core/subagents/researcher-subagent.agent.md index 9b3046d97..e6c6d2148 100644 --- a/.github/agents/hve-core/subagents/researcher-subagent.agent.md +++ b/.github/agents/hve-core/subagents/researcher-subagent.agent.md @@ -64,7 +64,7 @@ Initial chat response, emit at most: * 1 line: subagent file path (the parent re-reads this file when it needs detail). * 1 line: status (Complete / Blocked / Needs Clarification). * Up to 7 bullet-point key findings (each โ‰ค 240 chars). Prioritize findings the parent cannot act on without reading the file. -* A checklist of recommended next research not completed during this session. +* A checklist of up to 5 recommended next research items not completed during this session. * Up to 3 clarifying questions, only when blocking. * 1 short "Full Detail" pointer line: "Re-read `` for complete evidence, code blocks, file/line citations, and rejected alternatives." diff --git a/.github/agents/hve-core/subagents/rpi-validator.agent.md b/.github/agents/hve-core/subagents/rpi-validator.agent.md index fbc01caf0..7cb144651 100644 --- a/.github/agents/hve-core/subagents/rpi-validator.agent.md +++ b/.github/agents/hve-core/subagents/rpi-validator.agent.md @@ -70,7 +70,7 @@ Initial chat response, emit at most: * 1 line: review log file path (the parent re-reads this file when it needs detail). * 1 line: validation status (Pass / Fail). * Up to 7 bullet-point findings (each โ‰ค 240 chars). Prioritize schema violations and missing required sections. -* A checklist of recommended next validations not completed during this session. +* A checklist of up to 5 recommended next validations not completed during this session. * Up to 3 clarifying questions, only when blocking. * 1 short "Full Detail" pointer line: "Re-read `` for complete RPI artifact validation details." From 3a56ec0ac8c3a189125b1a2334c3328def78d3fc Mon Sep 17 00:00:00 2001 From: katriendg Date: Tue, 5 May 2026 07:23:39 +0000 Subject: [PATCH 9/9] refactor(agents): standardize response format language to lowercase for consistency MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit ๐Ÿ”ง - Generated by Copilot --- .github/agents/hve-core/rpi-agent.agent.md | 4 ++-- .../hve-core/subagents/implementation-validator.agent.md | 4 ++-- .github/agents/hve-core/subagents/plan-validator.agent.md | 4 ++-- .../agents/hve-core/subagents/researcher-subagent.agent.md | 4 ++-- .github/agents/hve-core/subagents/rpi-validator.agent.md | 4 ++-- .github/agents/hve-core/task-implementor.agent.md | 4 ++-- .github/agents/hve-core/task-planner.agent.md | 4 ++-- .github/agents/hve-core/task-researcher.agent.md | 4 ++-- .github/agents/hve-core/task-reviewer.agent.md | 4 ++-- 9 files changed, 18 insertions(+), 18 deletions(-) diff --git a/.github/agents/hve-core/rpi-agent.agent.md b/.github/agents/hve-core/rpi-agent.agent.md index 25f03ad5e..9a492f5a1 100644 --- a/.github/agents/hve-core/rpi-agent.agent.md +++ b/.github/agents/hve-core/rpi-agent.agent.md @@ -93,11 +93,11 @@ Treat the phase guidance below as operating defaults rather than ceremony. Deleg ## Context Discipline -After any subagent returns, this turn MUST be lean: +After any subagent returns, this turn must be lean: 1. Emit one compact line per subagent (subagent name + one-line outcome + tracking file path). 2. Update the relevant `.copilot-tracking/` file via a single edit if needed. -3. Stop. Do NOT re-read large planning, research, or details files in the closing turn. Do NOT re-quote subagent payloads. Do NOT narrate the next phase plan. +3. Stop. Do not re-read large planning, research, or details files in the closing turn. Do not re-quote subagent payloads. Do not narrate the next phase plan. Choose the lightest response mode that satisfies the request: diff --git a/.github/agents/hve-core/subagents/implementation-validator.agent.md b/.github/agents/hve-core/subagents/implementation-validator.agent.md index 2198701ca..626301ca8 100644 --- a/.github/agents/hve-core/subagents/implementation-validator.agent.md +++ b/.github/agents/hve-core/subagents/implementation-validator.agent.md @@ -199,7 +199,7 @@ During any validation scope, note issues that fall outside predefined categories ## Response Format -The subagent ALWAYS writes complete validation findings to the review log before returning. The chat response is an executive summary only. Full fidelity lives on disk. +The subagent always writes complete validation findings to the review log before returning. The chat response is an executive summary only. Full fidelity lives on disk. Initial chat response, emit at most: * 1 line: review log file path (the parent re-reads this file when it needs detail). @@ -208,4 +208,4 @@ Initial chat response, emit at most: * Up to 3 clarifying questions, only when blocking. * 1 short "Full Detail" pointer line: "Re-read `` for complete validation output, test results, and remediation guidance." -Do NOT paste full test output, lint dumps, or complete file diffs into the chat response. The review log is the source of truth. +Do not paste full test output, lint dumps, or complete file diffs into the chat response. The review log is the source of truth. diff --git a/.github/agents/hve-core/subagents/plan-validator.agent.md b/.github/agents/hve-core/subagents/plan-validator.agent.md index 52a943971..c4f15cbb7 100644 --- a/.github/agents/hve-core/subagents/plan-validator.agent.md +++ b/.github/agents/hve-core/subagents/plan-validator.agent.md @@ -94,7 +94,7 @@ When prior planning logs are available, cross-run comparison notes reference res ## Response Format -The subagent ALWAYS writes complete validation findings to the Planning Log before returning. The chat response is an executive summary only. Full fidelity lives on disk. +The subagent always writes complete validation findings to the Planning Log before returning. The chat response is an executive summary only. Full fidelity lives on disk. Initial chat response, emit at most: * 1 line: planning log file path (the parent re-reads this file when it needs detail). @@ -104,4 +104,4 @@ Initial chat response, emit at most: * Up to 3 clarifying questions, only when blocking. * 1 short "Full Detail" pointer line: "Re-read `` for complete discrepancy details, evidence, and recommended fixes." -Do NOT paste full discrepancy tables, complete plan excerpts, or research quotes into the chat response. The planning log is the source of truth. +Do not paste full discrepancy tables, complete plan excerpts, or research quotes into the chat response. The planning log is the source of truth. diff --git a/.github/agents/hve-core/subagents/researcher-subagent.agent.md b/.github/agents/hve-core/subagents/researcher-subagent.agent.md index e6c6d2148..8d27aa804 100644 --- a/.github/agents/hve-core/subagents/researcher-subagent.agent.md +++ b/.github/agents/hve-core/subagents/researcher-subagent.agent.md @@ -58,7 +58,7 @@ External URLs may still use markdown link syntax. ## Response Format -The subagent ALWAYS writes complete findings to its subagent file before returning. The chat response is an executive summary only. Full fidelity lives on disk. +The subagent always writes complete findings to its subagent file before returning. The chat response is an executive summary only. Full fidelity lives on disk. Initial chat response, emit at most: * 1 line: subagent file path (the parent re-reads this file when it needs detail). @@ -68,4 +68,4 @@ Initial chat response, emit at most: * Up to 3 clarifying questions, only when blocking. * 1 short "Full Detail" pointer line: "Re-read `` for complete evidence, code blocks, file/line citations, and rejected alternatives." -Do NOT paste file contents, code blocks, long quotes, or full evidence tables into the chat response. The subagent file is the source of truth. +Do not paste file contents, code blocks, long quotes, or full evidence tables into the chat response. The subagent file is the source of truth. diff --git a/.github/agents/hve-core/subagents/rpi-validator.agent.md b/.github/agents/hve-core/subagents/rpi-validator.agent.md index 7cb144651..561928def 100644 --- a/.github/agents/hve-core/subagents/rpi-validator.agent.md +++ b/.github/agents/hve-core/subagents/rpi-validator.agent.md @@ -64,7 +64,7 @@ Create and update the validation document progressively documenting: ## Response Format -The subagent ALWAYS writes complete validation findings to the review log before returning. The chat response is an executive summary only. Full fidelity lives on disk. +The subagent always writes complete validation findings to the review log before returning. The chat response is an executive summary only. Full fidelity lives on disk. Initial chat response, emit at most: * 1 line: review log file path (the parent re-reads this file when it needs detail). @@ -74,4 +74,4 @@ Initial chat response, emit at most: * Up to 3 clarifying questions, only when blocking. * 1 short "Full Detail" pointer line: "Re-read `` for complete RPI artifact validation details." -Do NOT paste full artifact contents, schema dumps, or long quotes into the chat response. The review log is the source of truth. +Do not paste full artifact contents, schema dumps, or long quotes into the chat response. The review log is the source of truth. diff --git a/.github/agents/hve-core/task-implementor.agent.md b/.github/agents/hve-core/task-implementor.agent.md index 76623b47c..51d844dde 100644 --- a/.github/agents/hve-core/task-implementor.agent.md +++ b/.github/agents/hve-core/task-implementor.agent.md @@ -58,11 +58,11 @@ Subagents can run in parallel when investigating independent topics or executing ## Context Discipline -After any subagent returns, this turn MUST be lean: +After any subagent returns, this turn must be lean: 1. Emit one compact line per subagent (subagent name + one-line outcome + tracking file path). 2. Update the relevant `.copilot-tracking/` file via a single edit if needed. -3. Stop. Do NOT re-read large planning, research, or details files in the closing turn. Do NOT re-quote subagent payloads. Do NOT narrate the next phase plan. +3. Stop. Do not re-read large planning, research, or details files in the closing turn. Do not re-quote subagent payloads. Do not narrate the next phase plan. Choose the lightest response mode that satisfies the request: diff --git a/.github/agents/hve-core/task-planner.agent.md b/.github/agents/hve-core/task-planner.agent.md index ddd773366..89e0b2a3d 100644 --- a/.github/agents/hve-core/task-planner.agent.md +++ b/.github/agents/hve-core/task-planner.agent.md @@ -60,11 +60,11 @@ Subagents can run in parallel when investigating independent topics or validatin ## Context Discipline -After any subagent returns, this turn MUST be lean: +After any subagent returns, this turn must be lean: 1. Emit one compact line per subagent (subagent name + one-line outcome + tracking file path). 2. Update the relevant `.copilot-tracking/` file via a single edit if needed. -3. Stop. Do NOT re-read large planning, research, or details files in the closing turn. Do NOT re-quote subagent payloads. Do NOT narrate the next phase plan. +3. Stop. Do not re-read large planning, research, or details files in the closing turn. Do not re-quote subagent payloads. Do not narrate the next phase plan. Choose the lightest response mode that satisfies the request: diff --git a/.github/agents/hve-core/task-researcher.agent.md b/.github/agents/hve-core/task-researcher.agent.md index 759fa674e..efdbbd831 100644 --- a/.github/agents/hve-core/task-researcher.agent.md +++ b/.github/agents/hve-core/task-researcher.agent.md @@ -52,11 +52,11 @@ Subagents can run in parallel when investigating independent topics or sources. ## Context Discipline -After any subagent returns, this turn MUST be lean: +After any subagent returns, this turn must be lean: 1. Emit one compact line per subagent (subagent name + one-line outcome + tracking file path). 2. Update the relevant `.copilot-tracking/` file via a single edit if needed. -3. Stop. Do NOT re-read large planning, research, or details files in the closing turn. Do NOT re-quote subagent payloads. Do NOT narrate the next phase plan. +3. Stop. Do not re-read large planning, research, or details files in the closing turn. Do not re-quote subagent payloads. Do not narrate the next phase plan. Choose the lightest response mode that satisfies the request: diff --git a/.github/agents/hve-core/task-reviewer.agent.md b/.github/agents/hve-core/task-reviewer.agent.md index 2202040e5..c5e231a91 100644 --- a/.github/agents/hve-core/task-reviewer.agent.md +++ b/.github/agents/hve-core/task-reviewer.agent.md @@ -39,11 +39,11 @@ Reviews completed implementation work from `.copilot-tracking/` artifacts. Valid ## Context Discipline -After any subagent returns, this turn MUST be lean: +After any subagent returns, this turn must be lean: 1. Emit one compact line per subagent (subagent name + one-line outcome + tracking file path). 2. Update the relevant `.copilot-tracking/` file via a single edit if needed. -3. Stop. Do NOT re-read large planning, research, or details files in the closing turn. Do NOT re-quote subagent payloads. Do NOT narrate the next phase plan. +3. Stop. Do not re-read large planning, research, or details files in the closing turn. Do not re-quote subagent payloads. Do not narrate the next phase plan. Choose the lightest response mode that satisfies the request: