feat(skills): require 5 RISA + 2 WILD floor in wild-risa-balance#37
Merged
naimkatiman merged 4 commits intomainfrom Apr 28, 2026
Merged
feat(skills): require 5 RISA + 2 WILD floor in wild-risa-balance#37naimkatiman merged 4 commits intomainfrom
naimkatiman merged 4 commits intomainfrom
Conversation
Tightens the recommendation-block contract so every list emitted under
this skill ships at least 7 items (exactly 2 WILD + at least 5 RISA).
Pre-existing prose ("1-2 bold items" / "the safe items") was loose
enough that lists could degrade to flat 2+2 blocks where the bold and
safe items competed equally and the bold ones lost by default.
- skills/wild-risa-balance.md: How-to-Apply rewritten with explicit
count floors; example expanded from 2+2 to 2+5 with annotated headers
- plugins/continuous-improvement/skills/wild-risa-balance/SKILL.md: same
edit mirrored so check-skill-mirror stays green
Captures the post-merge tightening of the recommendation contract so the plan doc remains the source of truth for the skill: 2 WILD + 5 RISA = 7 items minimum, with the file list and verification steps for the follow-up PR.
…tion Adds a wild-risa-floor test asserting that both the source skill and the plugin mirror contain the literal "2 WILD + 5 RISA = 7 items minimum", plus matching entries in the docs-substring inventory so the standalone lint catches drift before the broader test suite runs. - src/test/wild-risa-floor.test.mts: new test with two mirror checks - src/bin/check-docs-substrings.mts: 2 inventory entries appended - bin/, test/: regenerated build outputs (npm run build) The pair guards against future edits silently dropping the floor and re-degrading recommendation lists to flat 2+2 blocks.
…-the-recommendation Adds an "Upstream block-shape contract" line to Phase 1 (Pre-Flight) so the execution arm knows the expected block size when wild-risa-balance is upstream: 2 WILD + at least 5 RISA = 7 items minimum. Short or wrong-split blocks now get flagged to the user as a one-line note before the walk begins, rather than being walked silently. Mirrored into the plugin SKILL.md so check-skill-mirror stays green.
![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 7-item floor requirement for recommendation lists, mandating exactly 2 WILD items and at least 5 RISA items to prevent list degradation. The changes include documentation updates, cross-references in the execution skill, and new CI tests. A bug was identified in the repository root path calculation in the TypeScript test file.
| import { fileURLToPath } from "node:url"; | ||
|
|
||
| const __dirname = fileURLToPath(new URL(".", import.meta.url)); | ||
| const REPO_ROOT = join(__dirname, ".."); |
There was a problem hiding this comment.
The REPO_ROOT calculation is incorrect for a file located in src/test/. Since the skills/ and plugins/ directories are at the repository root, you need to go up two levels from this file's directory to reach the root.
const REPO_ROOT = join(__dirname, "..", "..");
naimkatiman
added a commit
that referenced
this pull request
Apr 28, 2026
Resolves conflicts in bin/check-docs-substrings.mjs and src/bin/check-docs-substrings.mts by keeping both the P-MAG literal-lock block (this branch) and the WILD/RISA floor literal-lock block (#37, just merged to main). Verified locally: - node bin/check-docs-substrings.mjs: 80/80 pass - node bin/check-skill-tiers.mjs: 13 tiered skills - node bin/check-skill-mirror.mjs: 13 pairs match - node --test test/past-mistake-gate.test.mjs: 8/8 pass - node --test test/wild-risa-floor.test.mjs: 4/4 pass - node --test test/skill-tiers.test.mjs: 17/17 pass
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
Follow-up to PR #35. Tightens the recommendation-block contract for
wild-risa-balanceso every list emitted under the skill ships at least 7 items (exactly 2 WILD + at least 5 RISA), and locks the contract via a CI assertion + a downstream cross-reference.Operator feedback after #35 merged: the loose prose ("1-2 bold items" / "the safe items") let lists degrade into flat 2+2 blocks where the bold and safe items competed equally and the bold ones lost by default. The 7-item floor restores the original intent — a real WILD/RISA contrast, not a flat menu.
Commits
feat(skills): require 5 RISA + 2 WILD floor in wild-risa-balance— How-to-Apply prose rewritten with explicit count floors; example expanded from 2+2 to 2+5 with annotated headers; plugin mirror updated.docs(plans): record 7-item floor amendment for wild-risa-balance— plan doc now carries the amendment so it stays the source of truth.feat(ci): lock 2 WILD + 5 RISA floor literal via docs-substring assertion— newwild-risa-floor.test.mts+ 2 new entries in the docs-substring inventory; both mirrors guarded.docs(skills): cross-reference wild-risa-balance floor in proceed-with-the-recommendation— the execution arm now flags blocks that arrive shorter than 7 instead of walking them silently.Test plan
node bin/check-skill-mirror.mjs— 13 pairs matchnode bin/check-docs-substrings.mjs— 74 assertions (was 72) all passnode bin/check-skill-tiers.mjs— 13 sources tier-taggednpm test— 242 tests pass, 0 fail