groove-work-audit — Claude Code skill: what it does, install
This skill checks the current branch for gaps and assigns each finding an action. Use it before merging changes or reviewing a focused topic.
npx skills add andreadellacorte/groove --skill groove-work-audit -a claude-codeWhat groove-work-audit does
groove-work-audit compares the current branch with main, reads the branch diff and related specifications, then looks for blind spots and missing work. Findings are grouped into three outcomes: fix immediately, prepare a specification, or create issues. Each finding includes severity and an effort estimate.
You can run the skill without arguments or give it a topic to narrow the review. The topic is sanitized before use by removing path separators, special characters, and traversal patterns such as ../. Specifications are read from .groove/memory/specs/. Analysis runs in an isolated general-purpose context; if that subagent fails, the skill falls back to a direct review.
The selected action is carried out: small fixes are applied directly, a specification is created with /groove-work-spec <topic>, or issues are created. The skill needs access to reading, writing, editing, search, git commands, and user questions.
What it helps with
- review the current branch against
main - spot gaps in code changes and related specifications
- sort findings by severity and effort
- apply trivial fixes during the review
- turn larger work into a specification or issues
Example prompts for the agent
- Review the current branch against
mainand find blind spots. - Focus the audit on
authenticationand group every finding by action. - Apply trivial fixes now and create issues for larger work.
How to install groove-work-audit
1Skills CLI (Vercel)
npx skills add andreadellacorte/groove --skill groove-work-audit -a claude-code2Manually: put the files in place
~/.claude/skills/groove-work-audit/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/groove-work-audit/
1Skills CLI (Vercel)
npx skills add andreadellacorte/groove --skill groove-work-audit -a codex2Manually: put the files in place
.agents/skills/groove-work-audit/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add andreadellacorte/groove --skill groove-work-audit -a cursor2Manually: put the files in place
.cursor/skills/groove-work-audit/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install andreadellacorte/groove groove-work-audit2Manually: put the files in place
.github/skills/groove-work-audit/SKILL.md1Skills CLI (Vercel)
npx skills add andreadellacorte/groove --skill groove-work-audit -a gemini-cli2Platform CLI
gemini skills install https://github.com/andreadellacorte/groove --path skills/groove-work-auditManually: put the files in place
.devin/skills/groove-work-audit/SKILL.mdThe legacy .windsurf/skills/ path still works
Skills CLI (Vercel)
npx skills add andreadellacorte/groove --skill groove-work-audit -a opencodeClaude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
Step by step, with commands for each platform: How to install skills in Claude Code: five ways
Questions about groove-work-audit
How do I install groove-work-audit in Claude Code?
Run npx skills add andreadellacorte/groove --skill groove-work-audit -a claude-code, then check ~/.claude/skills/groove-work-audit/SKILL.md.
What does groove-work-audit do?
It reviews the current branch against main, uses related specifications for context, and sorts findings into immediate fixes, specification work, or issues.
What happens if the subagent analysis fails?
The skill falls back to a direct review. You can still apply the selected action, create a specification with /groove-work-spec <topic>, or create issues.
Source text
SKILL.md · MIT
--- name: groove-work-audit description: "Review current branch for blindspots; fix now / needs spec / create issues." license: MIT allowed-tools: Read Write Edit Glob Grep Bash(git:*) AskUserQuestion metadata: author: andreadellacorte --- # groove-work-audit ## Outcome Blindspots and gaps are identified, actionable improvements are offered. ## Acceptance Criteria Findings are categorized by action: - **Fix Now**: Trivial effort, fix immediately - **Needs Spec**: Important, requires planning - **Create Issues**: Large effort or nice-to-have ## Constraints - Sanitize topic (if provided) for safe use — strip path separators, special characters, and traversal patterns (`../`) - Specs directory is always `.groove/memory/specs/` - Gather context: branch diff, relevant specs - Run analysis in isolated context (use general-purpose agent) - Fall back to direct review if subagent fails - Present findings with severity and effort estimates - Execute chosen action: fix directly, create spec via `/groove-work-spec <topic>`, or create issues ## Scope Options - No args: Review current branch vs main - Topic: Focus on specific area
© andreadellacorte · MIT · andreadellacorte/groove
