Help us improve
Share bugs, ideas, or general feedback.
From swe-skills
Audits a repository's code and docs to find important conventions, workflows, and architectural decisions that are missing from agent-facing guidance, then drafts review-ready updates. Use when a user says `capture repo knowledge`, `document implicit conventions`, `turn repo patterns into agent rules`, or asks what future agents should remember about a codebase. Do NOT use for a generic repo tour, onboarding walkthrough, or architecture summary that does not need reusable guidance updates.
npx claudepluginhub ckorhonen/swe-skillsHow this skill is triggered — by the user, by Claude, or both
Slash command
/swe-skills:capture-knowledgeThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Use this skill to convert implicit repo knowledge into explicit guidance that
Provides behavioral guidelines to reduce common LLM coding mistakes, focusing on simplicity, surgical changes, assumption surfacing, and verifiable success criteria.
Searches, retrieves, and installs Agent Skills from prompts.chat registry using MCP tools like search_skills and get_skill. Activates for finding skills, browsing catalogs, or extending Claude.
Provides UI/UX resources: 50+ styles, color palettes, font pairings, guidelines, charts for web/mobile across React, Next.js, Vue, Svelte, Tailwind, React Native, Flutter. Aids planning, building, reviewing interfaces.
Share bugs, ideas, or general feedback.
Use this skill to convert implicit repo knowledge into explicit guidance that future agents can follow reliably.
The job is not to summarize the repo. The job is to compare:
Then draft the missing entries and stop for review before writing anything.
Use this skill when the user wants to:
Do not use this skill for:
Before doing deep work, confirm or infer:
If the scope is missing, ask for the narrowest clarification needed before proceeding.
Inspect the current agent-facing guidance first, such as:
AGENTS.md, CLAUDE.md, and repo-level workflow docsCapture what is already documented so you do not restate it later.
Inspect the codebase, configs, tests, and representative implementations to understand how the repo is actually built and maintained.
Focus on repeated or operationally important patterns such as:
Compare observed repo behavior against the existing guidance and keep only the highest-signal gaps.
Prioritize:
Avoid:
For each proposed entry, include:
pattern, convention, or architectural-decisionGround every draft in concrete evidence from named files, directories, configs, or repeated code patterns. Distinguish observed facts from inferences.
Do not save proposed entries automatically.
Always:
If the user approves only a subset, save only that subset.
When writing approved entries, choose the narrowest correct destination:
Explain why that destination is the right place.
Produce a review packet with these sections:
Prefer 3-7 strong entries over a long weak list.
User says: Review this repo and capture the conventions future agents keep missing.
Actions:
AGENTS.md, README.md, and existing skillsResult: The user gets a small, high-signal review packet instead of a long repo summary.
User says: Turn the implicit release workflow in this repo into agent guidance, but do not save anything until I approve it.
Actions:
Result: Draft guidance is ready for review with a clear suggested destination.
If the existing guidance already captures the important behavior, say so
explicitly. Return No material knowledge gaps found rather than forcing weak
proposals.
Do not promote uncertainty into a rule. Present the item as a candidate for confirmation, explain what you observed, and state exactly what still needs to be verified.
If the user wants edits saved, you can prepare the write targets and draft text in the same pass, but still show the proposed entries before writing so the review step is explicit.