v0.4.0

Released: 2026-04-10

Canonical changelog: CHANGELOG.md

Highlights

• this release adds workspace ingress and mutation-gate techniques, audit-to-closeout proof loops, promotion-readiness surfaces, live technique receipt publishing, and antifragility/via-negativa guidance
• pinned validation evidence, repo/root scouting, and current-practice posture are strengthened while new shared-substrate and owner-sync techniques are promoted into the public set
• aoa-techniques remains a curated public corpus and documentation surface rather than a package or registry authority

Full Release Notes

Summary

• this release adds workspace ingress and mutation-gate techniques, audit-to-closeout proof loops, promotion-readiness surfaces, live technique receipt publishing, and antifragility/via-negativa guidance
• pinned validation evidence, repo/root scouting, and current-practice posture are strengthened while new shared-substrate and owner-sync techniques are promoted into the public set
• aoa-techniques remains a curated public corpus and documentation surface rather than a package or registry authority

Validation

• python scripts/release_check.py

Notes

• detailed corpus, generated-surface, report, and validation-asset coverage for this release remains enumerated below under Added, Changed, and Included in this release

Added

• workspace ingress and mutation-gate techniques plus audit-to-closeout
  proof-loop, recommendation-truth-vs-host-actionability, canonical-owner
  mirror, and pinned-validation techniques
• technique promotion-readiness surfaces and live technique receipt publishing
• antifragility recovery domains, via negativa techniques checklist, and quest
  feed validation surfaces

Changed

• strengthened pinned validation evidence, repo/root technique-kind scouting,
  and next-wave practice posture across the published corpus
• promoted new isolated shared-substrate and GitHub-only owner-sync techniques
  into the public set

Included in this release

• technique corpus growth across techniques/, docs/, generated/,
  reports/, config/, data/, and templates/, including the
  session-donor and session-harvest family, workspace ingress and
  mutation-gate techniques, audit-to-closeout proof loops, canonical-owner
  mirror, pinned validation, and live receipt publishing
• repo-local validation and release surfaces under .agents/, AGENTS.md,
  README.md, CONTRIBUTING.md, TECHNIQUE_INDEX.md, schemas/, scripts/,
  tests/, and quests/, including promotion-readiness manifests, via
  negativa guidance, quest-feed validation, and public corpus status refreshes

v0.3.0

Third public corpus release.

This changelog entry uses the release-prep merge date.

Summary

• 26 new public technique bundles since v0.2.0, growing the published corpus from 48 techniques to 74
• public corpus status is now 25 canonical techniques and 49 promoted techniques
• this release extends the corpus across handoff and continuation patterns, capability discovery and registry contracts, transcript lineage, fail-closed and approval-bound job control, OCR and media-ingest workflows, and Telegram normalization

Added

• AOA-T-0049 dependency-aware-task-graph, a promoted agent-workflows technique adapted from steveyegge/beads for explicit blocker graphs and derived ready work
• AOA-T-0050 ready-work-from-blocker-graph, a promoted agent-workflows technique adapted from steveyegge/beads for blocker-aware ready-frontier derivation
• AOA-T-0051 commit-triggered-background-review, a promoted agent-workflows technique adapted from roborev-dev/roborev for post-commit asynchronous review artifacts
• AOA-T-0052 review-findings-compaction, a promoted agent-workflows technique adapted from roborev-dev/roborev for findings verification and consolidation against current code
• AOA-T-0053 local-first-session-index, a promoted history technique adapted from wesm/agentsview for local searchable lookup over already-saved session artifacts
• AOA-T-0054 compaction-resilient-skill-loading, a promoted agent-workflows technique adapted from joshuadavidthomas/opencode-agent-skills for bounded post-compaction skill-availability recovery
• AOA-T-0055 requirements-design-tasks-ladder, a promoted agent-workflows technique adapted from gotalab/cc-sdd for a bounded requirement -> design -> task planning ladder
• AOA-T-0056 channelized-agent-mailbox, a promoted agent-workflows technique adapted from agentralabs/agentic-comm for durable named-channel communication with replay and explicit acknowledgment
• AOA-T-0057 structured-handoff-before-compaction, a promoted agent-workflows technique adapted from thebasedcapital/nightcrawler with supporting checkpoint framing from yan5xu/code-relay for explicit continuation packets before context compaction or rollover
• AOA-T-0058 receipt-confirmed-handoff-packet, a promoted agent-workflows technique adapted from jeremiah-k/agor with supporting explicit-acceptance surfaces from ax-platform/ax-platform-mcp for visible handoff receipt before continuation
• AOA-T-0059 git-verified-handoff-claims, a promoted agent-workflows technique adapted from thebasedcapital/nightcrawler with supporting snapshot-verification posture from jeremiah-k/agor for repo-backed verification of handoff claims before continuation
• AOA-T-0060 session-opening-ritual-before-work, a promoted agent-workflows technique adapted from thebasedcapital/nightcrawler for explicit pre-mutation session-start reading and baseline verification before resumed work begins
• AOA-T-0061 cross-repo-resource-map-bootstrap, a promoted agent-workflows technique adapted from yan5xu/code-relay for task-bounded cross-repo startup maps that name which repos and surfaces matter before continuation
• AOA-T-0062 episode-bounded-agent-loop, a promoted agent-workflows technique adapted from thebasedcapital/nightcrawler for checkpointed multi-episode continuation with explicit continue, stop, or escalate decisions
• AOA-T-0063 versioned-agent-registry-contract, a promoted docs technique adapted from agntcy/dir for named versioned registry-entry contracts with explicit references and bounded metadata
• AOA-T-0064 capability-discovery, a promoted docs technique adapted from agntcy/dir for bounded discovery-query contracts over already-published capability records
• AOA-T-0065 mcp-gateway-proxy, a promoted agent-workflows technique adapted from lasso-security/mcp-gateway for one reviewable proxy seam in front of configured MCP servers
• AOA-T-0066 transcript-replay-artifact, a promoted history technique adapted from es617/claude-replay with supporting context from wesm/agentsview for post-capture replay artifacts over saved sessions
• AOA-T-0067 transcript-linked-code-lineage, a promoted history technique adapted from git-ai-project/git-ai for bounded code-to-session provenance links
• AOA-T-0068 fail-closed-evidence-gate, a promoted agent-workflows technique adapted from Clyra-AI/gait for fail-closed execution gating with reviewable evidence output
• AOA-T-0069 approval-bound-durable-jobs, a promoted agent-workflows technique adapted from Clyra-AI/gait for durable jobs that pause and resume across an explicit approval seam
• AOA-T-0070 two-stage-document-ocr-pipeline, a promoted agent-workflows technique adapted from PaddleOCR and docTR for staged OCR handoff before later extraction or review
• AOA-T-0071 template-backed-field-extraction-after-ocr, a promoted agent-workflows technique adapted from invoice2data, receiptparser, and receipt-parser-legacy for bounded post-OCR field extraction through explicit templates, heuristics, and review fallback
• AOA-T-0072 perceptual-media-dedupe-with-threshold-review, a promoted agent-workflows technique adapted from imagededup and imgdupes for reviewable near-duplicate media grouping before later cleanup actions
• AOA-T-0073 semantic-media-bucketing-with-vision-plus-ocr, a promoted agent-workflows technique adapted from CLIP and PaddleOCR for confidence-aware mixed-media bucketing through bounded taxonomy and OCR side text
• AOA-T-0074 telegram-export-normalization-to-local-store, a promoted agent-workflows technique adapted from Telethon, TDLib, opentele, Chatistics, tg-archive, and telegram-mcp for resumable Telegram-source normalization into a provenance-preserving local store
• live questbook projection surfaces under generated/quest_catalog.min.json, generated/quest_dispatch.min.json, and matching example outputs
• downstream technique feed contracts and feat adjunct surfaces for current consumer layers

Changed

• promoted AOA-T-0028 confirmation-gated-mutating-action to canonical after GitHub Copilot's public coding-agent approval surfaces confirmed the same explicit confirmation-before-mutation seam beyond the donor lineage
• promoted AOA-T-0031 shell-composable-agent-invocation to canonical after OpenAI Codex CLI's public codex exec surface confirmed the same stdin/stdout/file-first one-shot shell contract beyond the donor lineage
• promoted AOA-T-0044 versionable-session-transcripts to canonical after claude-code-log confirmed a second public post-capture Markdown transcript-export surface beyond the donor product family
• promoted AOA-T-0053 local-first-session-index to canonical after coding-agent-search (cass) confirmed a second public local-first derivative session-index surface beyond the donor product family
• current corpus status is now 25 canonical techniques and 49 promoted techniques

Included in this release

• the current 74-bundle technique corpus under techniques/ plus the updated TECHNIQUE_INDEX.md
• questbook projection surfaces, downstream feed contracts, capsules, sections, checklists, examples, evidence notes, semantic reviews, and shadow reviews under generated/ and docs/

Validation

• python scripts/release_check.py

Notes

• this release remains a curated public corpus and validated documentation surface rather than a package or registry artifact

aoa-techniques v0.2.0

Second public corpus release.

This changelog entry uses the release-prep merge date.

Added

• 35 new public technique bundles since v0.1.0, growing the published corpus from 13 techniques to 48
• corpus coverage now spans 9 agent-workflows techniques, 24 docs techniques, 12 evaluation techniques, and the first 3 history techniques
• the first public KAG/source-lift family inside the docs domain, including AOA-T-0018 through AOA-T-0022
• the first bounded history domain for session and history artifacts that stay local-first and reviewable without widening into memory ownership, including AOA-T-0026, AOA-T-0044, and AOA-T-0045
• new repo-owned maintainer and navigation docs, including docs/START_HERE.md, docs/TECHNIQUE_SELECTION_GUIDE.md, docs/SEMANTIC_REVIEW_GUIDE.md, docs/EXTERNAL_IMPORT_RUNBOOK.md, docs/DONOR_REFINERY_RUBRIC.md, docs/LONG_GAP_CANON_DESIGN.md, docs/DEEP_AUDIT_ROADMAP.md, docs/EXTERNAL_TECHNIQUE_CANDIDATES.md, and docs/CROSS_LAYER_TECHNIQUE_CANDIDATES.md
• new derived surface families for technique capsules, repo-doc routing, technique sections, checklists, examples, evidence notes, GitHub review templates, semantic reviews, and shadow reviews

Changed

• public corpus status is now 21 canonical techniques and 27 promoted techniques, up from 9 canonical and 4 promoted in v0.1.0
• the canonical default set expanded across agent workflows, docs, evaluation, and KAG/source-lift surfaces, including AOA-T-0004, AOA-T-0013 through AOA-T-0019, AOA-T-0021, AOA-T-0023, and AOA-T-0034
• evidence and review posture is stronger across the corpus through broader second-context-adaptation, canonical-readiness, external-origin, external-import-review, and canonical-only adverse-effects-review coverage
• repo routing now centers on docs/START_HERE.md and the bounded pick -> inspect -> expand -> object use operating path
• release and validation posture now centers on python scripts/release_check.py, with tighter generator-drift checks, repo-doc and review-surface validation, broader public-hygiene URL scanning, and cleaner worktree behavior

Included in this release

• technique bundles under techniques/ plus the expanded TECHNIQUE_INDEX
• capsule surfaces: docs/TECHNIQUE_CAPSULES.md, docs/TECHNIQUE_CAPSULE_GUIDE.md, generated/technique_capsules.json, and generated/technique_capsules.min.json
• repo-doc routing surfaces: docs/REPO_DOC_SURFACES.md, generated/repo_doc_surface_manifest.json, and docs/REPO_DOC_SURFACE_LIFT_GUIDE.md
• source-lift reader and guide surfaces: docs/TECHNIQUE_SECTIONS.md, docs/TECHNIQUE_SECTION_LIFT_GUIDE.md, docs/TECHNIQUE_CHECKLISTS.md, docs/TECHNIQUE_CHECKLIST_LIFT_GUIDE.md, docs/TECHNIQUE_EXAMPLES.md, docs/TECHNIQUE_EXAMPLE_LIFT_GUIDE.md, docs/EVIDENCE_NOTE_SURFACES.md, and docs/EVIDENCE_NOTE_PROVENANCE_GUIDE.md
• review routing surfaces: docs/SHADOW_PATTERNS.md, docs/PUBLISHED_SUMMARY_SHADOW_REVIEW.md, docs/EVALUATION_CHAIN_SHADOW_REVIEW.md, generated/shadow_review_manifest.json, generated/semantic_review_manifest.json, and generated/github_review_template_manifest.json
• governance and intake surfaces under .github/ plus the release and validation helpers under scripts/

Validation

• python scripts/release_check.py
• the bounded release check reruns repo-doc, catalog, capsule, section, checklist, example, evidence-note, GitHub review-template, semantic-review, and shadow-review builders before unittest and validate_repo

Notes

• this release remains a curated public corpus and validated documentation surface rather than a package or registry artifact
• package publishing to PyPI, npm, or other registries remains out of scope for v0.2.0
• release identity for this repository remains the changelog entry, Git tag, and GitHub release body

aoa-techniques v0.1.0

aoa-techniques v0.1.0

First public baseline release of aoa-techniques.

Added

• initial public release of aoa-techniques as a public library of reusable techniques for coding agents and humans
• repository entry documents: README.md, AGENTS.md, CONTRIBUTING.md, SECURITY.md, and WALKTHROUGH.md
• repository-wide technique map in TECHNIQUE_INDEX.md
• curated public technique catalog containing:
  • 9 canonical techniques
  • 4 promoted techniques
• public templates, schemas, and validation helpers for technique authoring and promotion

Included in this release

• technique bundles under techniques/
• generated selection and semantic-review navigation surfaces referenced from README.md
• bounded KAG-oriented manifest pilot series for:
  • section manifests
  • checklist manifests
  • example manifests
  • evidence-note manifests
  • GitHub review template manifests
  • semantic review manifests

Validation

Documented local validation path:

• python -m unittest discover -s tests
• python scripts/validate_repo.py

Notes

• this is the first public baseline release for the repository
• package publishing to PyPI, npm, or other registries is out of scope for v0.1.0
• release emphasis is the curated public technique corpus and its repo-level validation/documentation surface