From dev-toolkit
Code-specific layer over the base git-workflow skill: it defers all git mechanics (branch/commit/PR) to git-workflow and adds the code-only gates — run tests/lint/build before pushing, keep CI green, conventional-commit scopes.
How this skill is triggered — by the user, by Claude, or both
Slash command
/dev-toolkit:code-developmentWhen to use
Any task that will result in committed product-code changes — "add a feature", "fix the bug", "refactor", "update the styling", "wire up the endpoint".
The summary Claude sees in its skill listing — used to decide when to auto-load this skill
A thin layer over the base **`git-workflow`** skill, for changes to product code. All the
A thin layer over the base git-workflow skill, for changes to product code. All the
git — branching, commits, pushing, opening the PR, recovering from blockers — lives in
git-workflow. This skill only adds the parts that are specific to code.
For every code task, in order:
agent/skills/git-workflow/SKILL.md (and its
references/git-workflow.md) to preflight, branch off the base, and — at the end — push
and open the PR. Do not edit committed files while still on the base branch.npm test / npm run lint / npm run build, composer test, phpcs, etc.). CI must be green before review — never push code you haven't
run the checks on, and never merge a red PR.feat(checkout): …, fix(blocks): …, refactor(api): …. (The base reference covers the
general commit format; this is the code convention on top.)git add path/to/file) so generated build output, vendored
files, and secrets never sneak into a commit.| Topic | Reference |
|---|---|
| All git mechanics (branch, commit, push, PR, recovery) | agent/skills/git-workflow/references/git-workflow.md |
Add code-specific references here as the project grows (testing conventions, code style, deploy process, …). Each new reference becomes another gate this skill can route to.
git-workflow's hard rules applies (never commit to the base branch, never
force-push without permission, never skip hooks, confirm before anything destructive).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.
Creates, edits, and optimizes skills for Claude Code, including drafting, evaluating with test prompts, iterating on performance, and improving skill descriptions for better triggering accuracy.
npx claudepluginhub refactco/claude-toolkit --plugin dev-toolkit