retro
Conduct a retrospective on a coding agent session: propose environment improvements ranked by severity. Invoked by the user after a session or post-merge.
- License 许可
- Apache-2.0
- 所属仓库
- r3dlex/skills
- 最近更新
- 2026-10-29
- Stars
- ★ 0
- Forks
- 0
01 前言 Front Matter
- name
- retro
- description
- Conduct a retrospective on a coding agent session: propose environment improvements ranked by severity. Invoked by the user after a session or post-merge.
03 SKILL.md 正文
The user has asked for a retrospective. You are suggesting improvements to the coding agent's environment to improve future runs. This is the post-merge closure step of a shipped goal: once a PR merges, ask whether a retro is wanted before moving on.
Steps
-
Load the
write-a-skillskill for the writing style guide, and followwrite-agent-docswhen a finding turns into an edit of an agent-facing document. -
Read the primary sources for the session, or the merged PR and its review thread. Sources may include session logs on this machine; default to the current session when the user does not specify one.
-
Look for candidates for improvement in these categories.
- Navigation: how easy was it for the agent to find the right files? Are there hidden dependencies between files? Would a navigation pointer make it easier? Use when the session took a long time to find a piece of information.
- Automated checks: are there automated checks that could catch errors the agent made? Linting, typing, tests, filesystem linters? Read the repo's own check command first (its
package.json/build-toollint/checkscripts, its CI workflow), so a check that already exists but sits unwired or silently broken is the finding, not a reinvention. A repo with no guardrail (no pre-commit hook and no CI job running its lint/typecheck/test command) is itself a finding: an un-linted repo is a standing missed opportunity, not a neutral default. Use when the agent made a mistake an automated check could have caught, or the repo has no guardrail at all. - Coding standards: should the reviewer agent be given a new rule to enforce? Should an existing rule be removed or clarified? Classify the violation first: a mechanical one (a fixed syntactic pattern, a banned API, an import shape, a file-location rule) gets a deterministic check, full stop: a custom rule in the repo's own linter, a new pre-commit hook, or a new CI job, whichever the repo's language and existing guardrail make cheapest. Default to building the check over writing the rule. Reserve
CODING_STANDARDS.mdfor genuine judgement calls (cross-file consistency, "matches the surrounding style," anything no guardrail could ever substitute for). Use when the reviewer agent failed to catch a mistake. - Global AGENTS.md: are there any steering instructions that should be moved to coding standards (or automated checks) instead? Use when the AGENTS.md file is particularly large - in the repo OR the user's global scope.
- Tool economy: did the agent make expensive tool calls that could be streamlined? Is there any custom tooling (CLIs, MCPs) that is particularly token-inefficient? Use when the agent made an expensive tool call.
- No-ops: look for instructions in steering files that don't modify the agent's behavior. Use when the steering files are large and unwieldy.
- Information access: look for opportunities to increase the agent's access to information. Teeing dev server logs, readonly access to third-party services. Use when a crucial piece of information was not available to the agent.
- Present these candidates to the user, in order of severity, and record accepted ones in the retro summary.
Reference
Implementation vs Review
All work goes through two stages: implementation and review. The implementation agent has the most context pressure — it explores, writes code, and debugs failures. The review agent has the least: it receives a diff and usually needs no exploration.
This means the review agent should be responsible for imposing coding standards, not the implementation agent.
Files
CLAUDE.md/AGENTS.md: pushed into the context window of any agent working in this repo. Use incredibly sparingly — usually only navigation pointers to other files.CODING_STANDARDS.md(if present): read during review, not implementation. Add navigation pointers to docs folders if the standards file grows past ~1,000 lines.- Docs: reference files pointed to by other files. Look for existing docs before writing new ones.
- Skills: the right home for reusable workflows, since the frontmatter description enters the agent's context. Follow
write-a-skill. - Glossary: use the repo's
CONTEXT.md(orGLOSSARY.mdwhen the upstream rename has landed) vocabulary exactly.
End by declaring knowledge kind learning, then reading
ai-catapult-init/modules/knowledge-publish.md and running its host-neutral producer step — publish when a registry exists, otherwise record unpublished: <reason>; never fail the skill because of the registry.
04 同仓库更多 Skills
skill
Fixture.
stable-skill
Fixture stable-skill.
experimental-skill
Fixture experimental-skill.
deprecated-skill
Fixture deprecated-skill.
karpathy-guidelines
Apply think-before-coding, simplicity, surgical-change, and goal-driven engineering rules. Use when planning, implementing, or reviewing code.
karpathy-guidelines
Apply think-before-coding, simplicity, surgical-change, and goal-driven engineering rules. Use when planning, implementing, or reviewing code.