feat(01KVS12K): A skill that helped a story reach done is credited as useful even when no READY-FOR-MERGE verdict event was recorded

[jackmcintyre]

What changed

A skill that helped a story reach done is credited as useful even when no READY-FOR-MERGE verdict event was recorded

This change delivers the following acceptance criteria:

• AC1: Given a skill.invoke whose story_id reached done/ (its manifest is in .flow/state/done/) but for which NO joined later READY-FOR-MERGE reviewer.verdict event exists in telemetry, When computeSkillEffectiveness runs, Then that invocation is counted as a useful fire (useful_fire_count includes it and effectiveness_ratio is above 0) via a declared done/-manifest read, rather than scoring 0.
• AC2: Given a skill.invoke whose story did NOT reach done/ and has no joined READY-FOR-MERGE verdict, When effectiveness is computed, Then it is NOT credited (a genuinely unhelpful skill still scores 0); and a skill.invoke that already joins a later READY-FOR-MERGE reviewer.verdict still counts as a useful fire exactly as today.

Why

As a maintainer relying on the retro to keep the skill set healthy, I want skill-effectiveness to credit a skill.invoke as a useful fire when its joined story reached done/, not only when a later READY-FOR-MERGE reviewer.verdict event exists in telemetry, so that the retro does not draft a false skill-retire/revise proposal against a skill that helped a story complete through a path that recorded no READY-FOR-MERGE verdict event (for example an auto-merge-gate or operator-merge completion).

How to check it yourself

1. Run vitest on plugins/flow/mcp-server/src/tools/tests/skill-effectiveness-attribution.int.test.ts — the AC1 test verifies a done-manifest-only invoke is credited (useful_fire_count: 1, effectiveness_ratio: 1). 2. Run vitest on plugins/flow/mcp-server/src/tools/tests/compute-skill-effectiveness.test.ts — the done/-manifest suite (AC1/AC2a/AC2b/ENOENT) verifies credit vs no-credit vs ENOENT posture via injected seams. 3. Confirm no existing tests regressed — all 24 tests in both files must pass.

Risk and blast radius

Risk tier: medium

The change is scoped to the files in the diff below. Review the diff to judge its actual blast radius — including any effect on shared state, data schemas, or authentication paths — before approving.

What is explicitly not covered: reviewer verification of this summary's accuracy is handled by a separate companion story.

Evidence

The pre-pull-request build-and-test gate passed before this pull request was opened. No pull request can be opened by the automated flow unless both pnpm build and pnpm test exit 0 in the developer's working directory.

Per-criterion covering checks:

• AC1 → plugins/flow/mcp-server/src/tools/__tests__/skill-effectiveness-attribution.int.test.ts (automated test)
• AC2 → plugins/flow/mcp-server/src/tools/__tests__/compute-skill-effectiveness.test.ts (automated test)

Story: native:01KVS12KCEYJTVPN84NMRAPTT4
Spec: .flow/native-stories/01KVS12KCEYJTVPN84NMRAPTT4.md
ACs:

• AC1: Given a skill.invoke whose story_id reached done/ (its manifest is in .flow/state/done/) but for which NO joined later READY-FOR-MERGE reviewer.verdict event exists in telemetry, When computeSkillEffectiveness runs, Then that invocation is counted as a useful fire (useful_fire_count includes it and effectiveness_ratio is above 0) via a declared done/-manifest read, rather than scoring 0.
• AC2: Given a skill.invoke whose story did NOT reach done/ and has no joined READY-FOR-MERGE verdict, When effectiveness is computed, Then it is NOT credited (a genuinely unhelpful skill still scores 0); and a skill.invoke that already joins a later READY-FOR-MERGE reviewer.verdict still counts as a useful fire exactly as today.

Credit skill.invoke as useful when its story reached done/ with no READY-FOR-MERGE verdict event