feat(skills): add wild-risa-balance Tier-2 lens for split recommendation lists#35
Merged
naimkatiman merged 8 commits intomainfrom Apr 28, 2026
Merged
feat(skills): add wild-risa-balance Tier-2 lens for split recommendation lists#35naimkatiman merged 8 commits intomainfrom
naimkatiman merged 8 commits intomainfrom
Conversation
![gemini-code-assist[bot]](https://avatars.githubusercontent.com/in/956858?s=60&v=4){kind=small}
There was a problem hiding this comment.
Code Review
This pull request introduces a new Tier-2 skill named wild-risa-balance, designed to help agents balance creative, disruptive ideas (WILD) with realistic, executable tasks (RISA). The changes include the creation of a detailed plan document, the skill definition files in both source and plugin directories, and updates to the respective README files to include the new skill in the Tier 2 table. Feedback was provided to clarify the instructions regarding operator interaction, specifically ensuring the agent explicitly asks the operator to select a WILD pilot before proceeding with recommendations.
| 1. **Top block — WILD pilots.** 1–2 bold items. Present every item; the operator picks at most one to actually run. | ||
| 2. **Bottom block — RISA baseline.** The safe items that ship now regardless of the WILD bet. | ||
| 3. Within each block, rank descending by impact. | ||
| 4. Once the list is composed, wait for the operator's "proceed" signal before invoking `proceed-with-claude-recommendation`. Never auto-trigger it. This skill only changes how the list is composed. |
There was a problem hiding this comment.
The instruction to 'wait for the operator's "proceed" signal' is slightly ambiguous when WILD pilots are involved. Since the rule in step 1 is that the operator picks 'at most one' WILD item, the agent should be explicitly prompted to ask for this selection. Without a specific selection, the proceed-with-claude-recommendation skill (which this skill pairs with) defaults to executing the entire list in order, which would violate the 'at most one' constraint for WILD items.
Suggested change
| 4. Once the list is composed, wait for the operator's "proceed" signal before invoking `proceed-with-claude-recommendation`. Never auto-trigger it. This skill only changes how the list is composed. | |
| 4. Once the list is composed, explicitly ask the operator to select which WILD pilot (if any) to include alongside the RISA baseline. Wait for this selection before invoking proceed-with-claude-recommendation. Never auto-trigger it. This skill only changes how the list is composed. |
…ion lists Pairs WILD (Wild/Imaginative/Limitless/Disruptive) generation with RISA (Realistic/Important/Specific/Agreeable) execution so multi-item recommendation blocks split into a top WILD pilot bet and a bottom RISA baseline. Composition only — execution still routes through proceed-with-claude-recommendation. - skills/wild-risa-balance.md: new Tier-2 source skill (94 lines) - skills/README.md: row added under Tier 2 — expert-mode add-ons - plugins/continuous-improvement/skills/**: regenerated bundle - docs/plans/2026-04-28-wild-risa-balance-skill.md: plan doc
ECC / Everything Claude Code is affaan-m's project branding. This repo adopts continuous-improvement as its own brand across 25 source files in 5 commits, leaving git history untouched.
Replaces borrowed ECC branding with continuous-improvement on the public-facing surfaces (README skill table source column + marketplace listing descriptions for workspace-surface-audit). Step 1 of 5 per docs/plans/2026-04-28-rebrand-from-ecc.md.
Replaces 28 ECC references with continuous-improvement across the workspace-surface-audit skill, command, and both plugin mirrors. Step 2 of 5 per docs/plans/2026-04-28-rebrand-from-ecc.md.
Replaces ECC references with continuous-improvement across safety-guard and gateguard, plus their plugin mirrors. Includes "ECC 2.0" -> "v2" and "the ECC hook" -> "the continuous-improvement hook". Step 3 of 5 per docs/plans/2026-04-28-rebrand-from-ecc.md.
Updates origin frontmatter field from ECC to continuous-improvement across tdd-workflow, verification-loop, strategic-compact, and their plugin mirrors. Step 4 of 5 per docs/plans/2026-04-28-rebrand-from-ecc.md.
Replaces remaining ECC references with continuous-improvement across proceed-with-claude-recommendation skill and command, wild-risa-balance, skills/README skill table source column, the wild-risa plan doc, and the four plugin mirrors. Step 5 of 5 per docs/plans/2026-04-28-rebrand-from-ecc.md. Repo is now clean of borrowed ECC branding (excluding git history and the rebrand plan doc itself).
…, refresh plugin mirror Folds the post-rebase identifier-rename fallout into the wild-risa publication stack rather than landing as a separate commit: - skills/wild-risa-balance.md (3 cross-refs to the renamed skill) - docs/plans/2026-04-28-wild-risa-balance-skill.md (3 cross-refs) - docs/plans/2026-04-28-rebrand-from-ecc.md (3 cross-refs) - plugins/continuous-improvement/skills/wild-risa-balance/SKILL.md (regenerated mirror, 3 cross-refs) - plugins/continuous-improvement/skills/README.md (regenerated mirror, picks up the prior ECC->continuous-improvement rebrand for workspace-surface-audit) No body-language genericization in this commit; that pass is deferred to a separate PR per the post-rename plan. Verified: npm test 238/238 passing; final straggler grep returns only intentional historical refs (CHANGELOG v3.3.0 + 2 historical SHA URLs in docs/testing/proceed-with-the-recommendation.TESTING.md).
naimkatiman
force-pushed
the
feat/wild-risa-skill
branch
from
e5d4316 to
8b0474c
Compare
6 tasks
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
skills/wild-risa-balance.md(94 lines): decision-framing lens that pairs WILD (Wild/Imaginative/Limitless/Disruptive) generation with RISA (Realistic/Important/Specific/Agreeable) execution, used to split multi-item recommendation blocks into a top WILD pilot bet and a bottom RISA baseline.proceed-with-claude-recommendation. No hook change. No edit to the 7 Laws card.plugins/continuous-improvement/skills/wild-risa-balance/regenerated bynpm run build. Tier 2 row added toskills/README.mdand the generated bundled README.docs/plans/2026-04-28-wild-risa-balance-skill.mdrecords goal, WILL build / WILL NOT build, verification steps, and explicitly defers Option B (require WILD/RISA tagging insideproceed-with-claude-recommendation) and Option C (require WILD/RISA split in the 3-section close hook).Why
Multi-item recommendation lists tend to flatten bold options against safe ones, and the bold options lose by default. This skill makes the split deliberate: present a clearly-labeled top WILD block with 1–2 bold pilots and a clearly-labeled bottom RISA block with the safe baseline that ships regardless. The operator gets a single bold bet to weigh against a baseline they already trust.
Process
Built using
superpowers:subagent-driven-development:disciplinecross-reference, operator-vs-agent ambiguity in step 1, vague handoff in step 4).Test plan
node bin/check-skill-tiers.mjsexits 0 — all 13 source skills declare a recognized tiernpm test— 238 pass / 0 failnpm run build— bundle regenerated cleanly, source/generated parity verifiedgit statusclean after commit; only the five expected paths touchedproceed-with-claude-recommendation)Out of scope (deferred)
skills/proceed-with-claude-recommendation.md(Option B)hooks/three-section-close.mjs(Option C)bin/script