From system
Create, update, and audit repo-specific docs/dev guidance from target repo evidence; cross-validate code patterns, anti-pattern examples, decisions, gaps, and AI coding rules.
How this skill is triggered — by the user, by Claude, or both
Slash command
/system:maintain-dev-docsThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Use this skill to set up or maintain a repository-specific `docs/dev` documentation system for correct code-writing guidance. The skill owns only the target repository's `docs/dev` documents.
agents/openai.yamlassets/docs-dev/README.mdassets/docs-dev/ai.README.mdassets/docs-dev/architecture.README.mdassets/docs-dev/decisions.README.mdassets/docs-dev/domain.README.mdassets/docs-dev/engineering.README.mdassets/docs-dev/evolution.README.mdreferences/audit-evidence.mdreferences/decision-records.mdreferences/docs-dev-structure.mdreferences/example-validation.mdreferences/pattern-research.mdreferences/repo-scan.mdreferences/workflow.mdUse this skill to set up or maintain a repository-specific docs/dev documentation system for correct code-writing guidance. The skill owns only the target repository's docs/dev documents.
Do not make target-repo docs/designs, docs/brainstorms, docs/solutions, traceability.md, or completion-audit.md default outputs. If the work needs those artifacts, report that a separate requirements/design workflow is needed.
references/repo-scan.md.docs/dev is missing or empty, scaffold only neutral README skeletons from assets/docs-dev/.architecture/, engineering/, or domain/; write future proposals under evolution/.docs/dev.Select only the modes needed for the current request.
init: docs/dev is absent or empty. Read references/docs-dev-structure.md and use the neutral assets.baseline: The user asks to understand or document the repo's current stack, structure, or conventions. Read references/repo-scan.md.pattern-research: The user asks for best patterns, anti-patterns, examples, design patterns, or code-writing guidance. Read references/pattern-research.md and references/example-validation.md.dependency-change: Dependencies, package files, lockfiles, APIs, or external SDKs changed. Use repo scan plus pattern research for the changed dependency only.architecture-change: Module boundaries, public APIs, import direction, folder structure, or runtime boundaries changed. Read references/docs-dev-structure.md and references/decision-records.md.domain-change: Domain terms, business rules, states, or invariants changed. Update docs/dev/domain and relevant gaps.audit: The user asks whether docs/dev is stale, incomplete, inconsistent, or valid. Read references/audit-evidence.md.decision: A durable code-writing or documentation decision is made or superseded. Read references/decision-records.md.ai-rules: AI reading order, uncertainty handling, or coding-agent rules need updates. Update docs/dev/ai.references/workflow.md for non-trivial work.references/repo-scan.md before deriving repo-specific guidance.references/docs-dev-structure.md before creating or moving docs/dev files.references/pattern-research.md and references/example-validation.md before writing best/anti-pattern examples.references/decision-records.md before adding or superseding decisions.references/audit-evidence.md before claiming completion or auditing existing docs.Continue with a documented default for reversible local choices such as whether to keep a topic in a folder README or split it into a detailed file.
Ask before:
docs/dev;Allowed default outputs:
docs/dev/README.mddocs/dev/architecture/**docs/dev/engineering/**docs/dev/domain/**docs/dev/evolution/**docs/dev/ai/**docs/dev/decisions/**Not default outputs:
docs/designs/**docs/brainstorms/**docs/solutions/**traceability.mdcompletion-audit.mdIf completion evidence is needed, write it into docs/dev root or folder READMEs, docs/dev/ai/README.md, or docs/dev/decisions/README.md.
npx claudepluginhub nwleedev/system --plugin systemGuides creation, editing, and verification of skills for AI coding agents using test-driven development with subagent scenarios. Use when authoring or debugging skills.