- Home
- Skills
- Code Quality & Review
- Pre-push-qa
More screenshots
Works with the AI Agents you already use
Pre-push-qa
Automated pre-push QA suite for linting, type-checking, and security scanning across multiple languages.
$5
Pre-push-qa
Example session with this skill installed
Run a full pre-push QA on my current staged changes. This is a TypeScript + Python + Node monorepo. I want lint, types, tests, build, secret scanning and dependency audit before I push.
- Read your context and instructions
- Compiled the pre-push-qa
I ran the full pre-push-qa pipeline against your staged changes. Detected stacks:
Node.js (TypeScript) and
Python. 6 files staged across src/, api/ and lib/.
QA Verdict: ❌ FAIL
| Step | Tool | Status | Details |
|---|---|---|---|
| Diff Analysis | git | OK | 6 files staged for commit |
| Lint (TS) | eslint | PASS | 0 errors, 12 warnings |
| Type-check (TS) | tsc | FAIL | 2 type mismatches in src/auth.ts |
| Lint (Py) | ruff | PASS | Clean |
| Type-check (Py) | mypy | PASS | Clean |
| Static Analysis | semgrep | WARN | 1 low-confidence finding in api/handlers.py |
| Tests (TS) | jest | PASS | 47/47 passed in 3.2s |
| Tests (Py) | pytest | PASS | 22/22 passed in 1.8s |
| Build | tsc --build | FAIL | Blocked by type-check failure |
| Secrets | gitleaks | FAIL | AWS access key detected in api/config.py:14 |
| Deps (Node) | npm audit | WARN | 2 moderate CVEs in lodash, axios (non-blocking) |
| Deps (Py) | pip-audit | WARN | 1 low CVE in requests (non-blocking) |
| Patterns | grep-E | FAIL | subprocess shell=True at api/runner.py:87 |
Blocker 1 — Type error in src/auth.ts
src/auth.ts:42:12 - error TS2339: Property 'user_id' does not exist on type 'Session'.
src/auth.ts:89:5 - error TS2322: Type 'string | null' is not assignable to type 'string'.
Connects securely to your tools. The creator never sees your data.
About this skill
Manual pre-push checks are slow and error-prone, yet pushing broken code or leaked secrets destroys CI pipelines and risks production incidents. This skill provides a complete pre-push quality gate that runs lint, type-check, static analysis, tests, build, secret scanning, dependency audit and dangerous pattern detection — before your code leaves the machine.
What it does
- Diff analysis — scans staged changes via
git diff --cachedto target only relevant files. - Stack detection — identifies Node.js, Python, Go, Rust, and PHP environments automatically.
- Lint execution — runs eslint, ruff, golangci-lint, clippy, or phpcs based on detection.
- Type checking — verifies code integrity with tsc, mypy, phpstan, or compiler checks.
- Static analysis — semgrep, bandit, gosec for security-focused patterns.
- Targeted tests — pytest, npm test, go test, cargo test, phpunit for the modified scope.
- Build verification — ensures the project still compiles or bundles correctly.
- Secret scanning — gitleaks, trufflehog, detect-secrets catch leaked API keys, tokens, and passwords before they hit remote history.
- Dependency audit — npm audit, pip-audit, cargo audit report known CVEs as warnings (non-blocking).
- Dangerous pattern detection — flags eval, exec, subprocess shell=True, innerHTML, document.write, and other injection risks.
- Verdict reporting — clear PASS / WARN / FAIL status with a prioritized list by severity.
- Auto-fixes —
--fixapplies eslint --fix, ruff --fix, phpcs --fix and re-runs the affected steps.
How it works
- Run
bash scripts/install-hook.shto register the pre-push hook automatically. - Or run
bash scripts/run.shmanually to audit the current staged diff. - On Windows:
powershell -ExecutionPolicy Bypass -File scripts/run.ps1. - Every run logs to
.hermes/pre-push-qa.log(timestamp, verdict, duration, impacted files).
Options
--fix— apply automatic corrections--staged-only— check only staged files (default behavior)--no-security— skip secret/dependency/pattern scans--strict— promote all WARN to FAIL--json— additional JSON output for CI integration
Frameworks & tools
Supports Node.js, Python, Go, Rust, PHP, with basic Java and Ruby support. Requires Git and Bash 4+. Works natively on Linux/macOS, or via Git Bash/WSL on Windows. Compatible with Claude Code, Cursor, Codex CLI, Gemini CLI, GitHub Copilot, Windsurf, OpenClaw, Aider, Cline, VS Code and Grok.
Why this beats prompting it yourself
Writing a reliable cross-language pre-push gate requires diff parsing, per-tool detection, timeout handling, graceful fallbacks for missing binaries, and a coherent verdict aggregator. This skill handles all the boilerplate and edge cases — you get a consistent PASS/WARN/FAIL without writing 400 lines of bash.
Use cases
- Block git pushes that contain lint errors, type mismatches, or leaked credentials.
- Verify staged changes across Node, Python, Go, Rust and PHP stacks in one pass.
- Audit code quality locally before opening a Pull Request.
- Enforce security hygiene in monorepos with heterogeneous toolchains.
- Replace fragmented pre-commit setups with a single, portable gateway.
Known limitations
Requires Bash 4+; Windows users must use Git Bash or WSL. Missing tools are skipped with a warning rather than installed. Dependency audits are non-blocking by default (WARN, not FAIL) to avoid false-positive noise.
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.
Requires Bash 4+. Windows requires Git Bash or WSL. Skips missing tools without installing them — no auto-install of linters, type-checkers, or test runners. Dependency audits (npm audit, pip-audit, cargo audit) return WARN, not FAIL, to avoid CI noise. Does not replace a full CI pipeline; it is a local first-line defense.
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 1 month ago
- Software
- eslint
Node.jstsc
Python