More screenshots

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIDeepSeekDeepSeekMuseMuse+22 more

    Commit Message Doctor — Make Your Git History Reviewable

    1

    Reads a real git log, scores the commits that will confuse reviewers, and drafts the changelog grouped into features, fixes and breaking changes.

    Free

    Security scanned
    Commit Message Doctor — Make Your Git History Reviewable

    Commit Message Doctor — Make Your Git History Reviewable

    Example session with this skill installed

    git log -n 50 --format='%H%x09%s%x09%b%x1e' > commits.txt
    python3 scripts/commit_lint.py commits.txt --conventional --out review.md

    commit abc00001
    Update auth

    commit abc00002
    feat(api): add a webhook endpoint that verifies the signature and retries delivery with exponential backoff

    • Read your context and instructions
    • Compiled the commit message doctor

    commit-message-doctor 30 commits
    score 24/100 grade F findings 22
    high 9 medium 10 low 3

    CHANGELOG DRAFT
    Features (7)

    • Add idempotency keys to refund endpoint
    • Rank title matches above body matches

    Fixes (6)

    • Reject expired session tokens
    • Use captured amount for partial refunds

    Breaking changes

    • Require an idempotency key on refunds

    HIGH C6 abc00001 Update auth
    No Conventional Commits type on a repository that uses them.
    HIGH C3 abc00002 feat(api): add a webhook endpoint that verifies...
    Subject is 107 characters; GitHub truncates past 72.

    VERDICT: REWRITE BEFORE RELEASE

    Connects securely to your tools. The creator never sees your data.

    About this skill

    A branch can contain thirty commits and only a few sentences a reviewer can use. "Update auth", "fixed bug", three duplicates from a squash, and a breaking change hidden in lowercase: the history is the one review context that survives the pull request.

    This reads the log you already have and reports what a reviewer will struggle with:

    • Vague subjects such as "update", "fix stuff" and "WIP"
    • Past-tense and gerund openings that should be imperative
    • Subjects over 50 characters, with a hard flag past 72 where GitHub truncates
    • Missing Conventional Commits types, when your repository uses them
    • Empty scopes and lowercase breaking change: footers that parsers miss
    • Duplicate subjects left by squash or cherry-pick
    • Body lines over 72 columns

    Every run also drafts a changelog grouped into Features, Fixes, Performance, Refactors, Docs, Tests, Build, CI, Chores and Breaking changes. It is assembled from the subjects; a human still edits the release note.

    The shipped history is 30 commits with 22 findings: nine high, ten medium, three low. It includes a 107-character subject, a malformed feat():, a lowercase breaking footer and a duplicated fix. Score 24/100, exit 1. The clean history is 12 commits with no findings and a complete changelog draft.

    Good to know: this checks style and structure, not whether a subject describes the right change. Conventional Commits is opt-in because many repositories do not use it. The imperative check only recognises common past-tense and gerund openings; unusual verbs pass. It reads the log you give it, so a shallow clone only audits the commits in that clone. The changelog has no issue links, upgrade steps or security wording.

    Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Code Quality & Review.

    Audits structure/style only, not code intent. Requires Python 3. Recognizes common verbs for imperative checks; unusual verbs may pass. Does not generate issue links or migration steps.

    How to use

    1. 1

      Connect your agent to the Agensi MCP

      Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup

    2. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before buying a skill.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill goes through the trust checks below.

    Security scanned

    Checked by our 8-point scan before listing

    Fresh listing

    Recently published to Agensi

    Free to use

    No account required to browse

    Trust & safety

    Security scan

    Security scan passed 1 day ago

    • Free with an Agensi account

    Software
    Git

    Permissions

    Terminal / Shell
    Read Files
    Write Files
    Network Access
    Environment Variables
    Listed1 day ago

    What's inside

    Frequently Asked Questions

    This skill identifies non-compliant formatting, flags vague descriptions like "fix," and enforces the imperative mood to ensure your history is compatible with automated semantic versioning tools.

    The package includes the core Python auditing engine, a set of customizable linting rules, and a template for generating structured markdown changelogs.

    The tool is designed for any environment that can execute Python 3 scripts, making it compatible with AI agents that have terminal access or local file system capabilities.

    Yes, the script provides standard exit codes (0 for pass, 1 for fail), allowing you to block builds or pull requests if the commit history fails to meet your quality threshold.

    No, for security and privacy, the skill audits a local text export of your git log, meaning it never requires direct access to your remote repository credentials.

    While it checks for structural compliance and grammar (like the imperative mood), it does not verify if the code changes actually match the description in the message.

    New in Code Quality & Review

    Free