From dev
Implement changes following project conventions. Reads AGENTS.md and area-specific docs, then guides implementation, quality checks (pre-commit, tests), and commit with proper message format.
How this skill is triggered — by the user, by Claude, or both
Slash command
/dev:codeThis skill is limited to the following tools:
The summary Claude sees in its skill listing — used to decide when to auto-load this skill
Implement code changes following project conventions.
Implement code changes following project conventions.
Read the project's AGENTS.md for universal conventions, then load area-specific
docs as identified in /start.
Follow AGENTS.md conventions. Common rules:
Pre-commit hooks run automatically on git commit. To run manually:
bash .claude/scripts/format-and-lint.sh # staged files
bash .claude/scripts/format-and-lint.sh --all-files # all files
Run relevant tests per the project's test commands (check AGENTS.md for make targets).
Format:
<subsystem>: <what changed> (<TICKET-KEY>)
<why this change was made>
Pre-commit hooks run on commit — fix any failures and re-commit.
Do not stop after committing. Invoke the /pr skill immediately to create the pull request. The full workflow is a single uninterrupted pipeline:
/code → /pr → /poll <PR#>
Only pause if there is a genuine blocker that requires a decision from the user.
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.
npx claudepluginhub quay/ai-helpers --plugin dev