Works with the AI Agents you already use
Claude Code
Cursor
Codex CLI
DeepSeek
Muse+22 moreClaude CodeCursorCodex CLIDeepSeekMuse
Dots
Manus
Grok
Grok Bot+18 agentsPython CommitComposer — AST Conventional Commit Guard
Enforces strict Conventional Commits for AI agents and blocks hardcoded secrets before git staging.
$19.99
Python CommitComposer — AST Conventional Commit Guard
Example session with this skill installed
I updated the token authentication logic and added an integration test. Commit these staged changes and write a pull request description. Note: I accidentally left a .env file staged in the git index.
- Read your context and instructions
- Compiled the python commitcomposer
"audit_status": "BLOCKED",
"findings": [
{
"id": "f1",
"file": ".env",
"line": 0,
"severity": "HIGH",
"signature": "ERR_PRIVACY_BLOCK",
"message": "Potential secret file '.env' detected in staged changes. Commit aborted.",
"snippet": "git staged index"
},
{
"id": "f2",
"file": "commit_msg",
"line": 1,
"severity": "MEDIUM",
"signature": "ERR_LAZY_SUBJECT",
"message": "Initial intent string was too ambiguous for a professional commit subject.",
"snippet": "updated auth logic"
}
],
"suggested_commit": "fix(auth): prevent token expiry during long-running requests\n\nThe token refresh was previously tied to a static wall clock. This change anchors the refresh logic to the request start time to ensure stability during high-latency operations.\n\nCloses #142",
"metadata": {
"version": "1.0.0",
"timestamp": "2026-10-15T12:00:00Z",
"files_scanned": 2,
"security_status": "VIOLATION_DETECTED"
}
}
Connects securely to your tools. The creator never sees your data.
About this skill
Stop tolerating unreadable, chaotic repository history filled with lazy "fix", "update", and "wip" commit messages generated by local AI development sessions. Python CommitComposer is a deterministic static validation engine that forces AI coding agents to comply with rigid Conventional Commit specs while serving as an air-gapped pre-commit security gate.
Standard text-based system prompts degrade during extended context sessions, causing automated tools to gradually drop compliance rules. This skill wraps the prompt interface in a strict, native Python schema validator that exits with non-zero codes, physically stopping non-compliant or compromised commits from reaching your remote history.
What it does
- Conventional Commit Enforcement: Intercepts and validates formatting layouts against strict B2B type tokens (
feat,fix,refactor,test,chore,perf). - Context Drift Mitigation: Prevents AI agents from lapsing into lazy context definitions, forcing a 72-character imperative text layout boundary.
- Pre-Commit Privacy Interception: Deep-scans active staging indexes to detect and explicitly block high-severity exposures like un-gitignored
.envfiles, server credentials, and infrastructure configurations. - Automated Pull Request Synthesizer: Compiles atomic commit diff metadata into structured, professional Pull Request logs and markdown CHANGELOG updates.
- Machine-Readable Telemetry: Emits a schema-validated JSON payload optimized for native Git hooks, enterprise CI/CD verification pipelines, and local IDE buffers.
Detected Violation Signatures
- ERR_PRIVACY_BLOCK: Active staging area contains sensitive operational parameters,
.envdeployment sheets, or hardcoded API keys. - ERR_LAZY_SUBJECT: AI agent generated an ambiguous or descriptive-less header layout missing structural engineering depth.
- ERR_INVALID_SCOPE: Commit metadata scope flags parameters outside the defined configuration directory schema.
- REGEX_MISMATCH: Header structural failure missing the mandatory type-bracket compliance layout configuration.
When to use
- Standardizing repository version histories across distributed engineering environments using Cursor, Claude Code, or GitHub Copilot.
- Integrating automated, zero-hallucination pre-commit security gates to insulate codebases from credential leakage.
- Scaling distributed Python microservices where clear change logs are mandatory for automated version tagging.
When NOT to use
- Working outside local repository trees (requires a valid, active
.gitworkspace infrastructure block). - Technology tracks adopting Gitmoji templates or highly customized, non-standard commit format schemas.
- Executing active remote repository code writing modifications independently (this linter is strictly a passive validation gate).
Under the Hood & Security
- 100% Local & Air-Gapped: Operates entirely via standard Python stdlib abstractions (
file_system_write: false). Zero outbound cloud lookups guarantee corporate schemas, file patterns, and commit text structures never leak. - Contract-Tested Stability: Backed by a rigorous 12-point testing harness validating schema compliance boundaries, edge cases, and absolute zero-hallucination execution.
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.
Static validation constraints only. Operates without executing dynamic binary routines or third-party code compiling tools. Does not support non-standard layouts or formats built entirely on Gitmoji arrays. To maintain data isolation parameters, the tool reads file names and commit metadata variables without dynamically parsing or matching raw diff contents. Scans are automatically bypassed on repositories under 300 total lines of code.
How to use
- 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
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 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
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 29 days ago
- Software
Python
Git
Permissions
Allowed Hosts
File Scopes
Read-only access to local workspace file names and staging index logs. Evaluates structures completely offline with absolute zero filesystem mutation, runtime shell access, or external network vectors.