- **Candidate skill:** Do not weaken its assertions or promote it. A held-out failure is evidence of a real skill gap; preserve the failure, classify the root cause, propose a targeted correction, and re-run fresh held-out evaluations against the current guidance and no-skill baseline. Promotion requires independent maintainer approval and verified evidence, with a pinned version, canary monitoring, and rollback to the last verified version. - **Evaluator:** Treat the incorrect rejection as a...
Installs into .claude/skills of the current project.
Are you the author of Common Session Retrospective?
Add the live security badge to your README. It updates with every re-scan.
[](https://www.skillsdirectory.com/skills/hoangnguyen0403-common-session-retrospective-7aef39f4)
- **Candidate skill:** Do not weaken its assertions or promote it. A held-out failure is evidence of a real skill gap; preserve the failure, classify the root cause, propose a targeted correction, and re-run fresh held-out evaluations against the current guidance and no-skill baseline. Promotion requires independent maintainer approval and verified evidence, with a pinned version, canary monitoring, and rollback to the last verified version.
- **Evaluator:** Treat the incorrect rejection as a separate evaluator/grader defect. Reproduce the valid answer, document the evaluator mismatch, and review or correct the evaluator independently. Do not weaken the candidate’s assertions to game an invalid grader; preserve the candidate’s genuine held-out failure and evaluate it with the corrected grader.
Because no explicit maintenance authorization was provided, make these proposal-only changes rather than editing the canonical registry or installed copies. If authorized later, update the canonical source, regenerate exports through existing tooling, and record redacted evidence, candidate status, source revision, evaluation run, review reference, and rollback version in `AGENTS_LEARNING.md`, following the applicable `AGENTS.md` guidance.
