メニュー
Use when a plugin, skill, prompt, tool, command, or agent behavior was surprising, wrong, partial, unsafe, brittle, or worth preserving as a future eval fixture; also use when the user asks to report, submit, capture, or file an eval case.
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
SOC 職業分類に基づく
Use when preparing local changes, a branch, pull request, merge request, or merge-to-main for final review before publishing, opening a PR, merging, or claiming readiness.
Use when implementing any feature, bugfix, behavior change, or refactor before writing production code.
Use when the user asks to shape fuzzy planning, product/design/engineering tradeoffs, scope/spec/ticket plans, or says "challenge this/help me think"; delegate to an advisor subagent. Good for rough ideas, second opinions, load-bearing decisions, or scope cuts. Skip scoped implementation, narrow bugs, reviews, and "just build" requests.
Use when setting up a project-local eval harness for prompts, agents, judges, RAG, tools, or other LLM behavior, especially when the default should be free/OSS promptfoo with repo-owned artifacts.
Use when a pull request or merge request needs to be driven to merge — watching CI, responding to review feedback, and merging once green and approved, across GitHub, Forgejo, or GitLab.
Use when the user wants repository task tracking, shared agent task state, cross-worktree coordination, tiber setup/install/scaffold guidance, or task create/list/show/prioritize/validate/close workflows. Plugin install and session start are non-mutating; setup integration starts with dry-run previews.
| name | submit-eval-case |
| description | Use when a plugin, skill, prompt, tool, command, or agent behavior was surprising, wrong, partial, unsafe, brittle, or worth preserving as a future eval fixture; also use when the user asks to report, submit, capture, or file an eval case. |
Use this skill to turn observed AI-assistant behavior into a sanitized eval-case
issue for jwilger/ai-plugins.
Offer to submit an eval case when:
evals/fixtures/ as pass, fail, partial,
or adversarial coverage.Do not interrupt every ordinary bug fix. Offer only when the scenario is reusable or diagnostic.
Summarize the candidate eval case in neutral terms.
Scrub and anonymize the sanitized input, actual behavior, expected behavior,
and artifacts using references/scrubbing.md.
Classify the expected eval outcome: pass, fail, partial,
adversarial, or unsure.
Draft the issue title and body locally.
Show the preview to the user and ask for explicit approval before posting.
If approved, post with:
gh issue create \
--repo jwilger/ai-plugins \
--title "[eval-case]: <short title>" \
--label eval-case \
--label needs-triage \
--body-file <sanitized-body-file>
Report the created issue URL.
gh is unavailable or unauthenticated, leave the sanitized issue body in a
local temp file and give the exact gh issue create command.