Github Security Scanner Hook

GitHub Security Scanner Hook

Scan staged files for secrets and code-level security vulnerabilities before every git commit. Works with any language or file type — detect-secrets finds hardcoded credentials and API keys; semgrep finds injection flaws, XSS, path traversal, and OWASP Top 10 patterns across Python, JavaScript, Go, Ruby, Java, and more. Can run as a one-off manual scan or install as an automated git pre-commit hook.

Prerequisites

• Python 3.8+ available on PATH (to create the .venv)
• Git repository (any language — Python, JS, Go, Ruby, Java, etc.)
• Staged files to scan, or --full-repo to scan everything
• Internet access on first run only (to pip-install detect-secrets and semgrep)
• Write access to .git/hooks/ (for hook installation only)

Workflow

Step 1 — Set Up .venv and Install Tools

Always use .venv — create it if it does not exist, then resolve the correct interpreter:

[ -d .venv ] || python -m venv .venv
PYTHON=$(if [ -f .venv/Scripts/python ]; then echo .venv/Scripts/python; else echo .venv/bin/python; fi)

Install detect-secrets and semgrep into the venv if not already present:

$PYTHON -m pip show detect-secrets > /dev/null 2>&1 || $PYTHON -m pip install detect-secrets
$PYTHON -m pip show semgrep > /dev/null 2>&1 || $PYTHON -m pip install semgrep

Both tools are optional independently — if semgrep is not installed the scan still runs with detect-secrets only. Semgrep uses the bundled rules in rules/security.yml (no login or network access required).

Step 2 — Run Security Scan on Staged Files

Run scan-staged.py to check all currently staged files:

$PYTHON .github/skills/github-security-scanner-hook/scripts/scan-staged.py

To scan the entire repository instead of only staged files:

$PYTHON .github/skills/github-security-scanner-hook/scripts/scan-staged.py --full-repo

The script prints a JSON summary to stdout and exits with:

• Exit code 0 — no findings (clean)
• Exit code 1 — one or more HIGH severity findings (would block the commit)
• Exit code 2 — only MEDIUM or LOW severity findings (warnings, no block)

Step 3 — Review and Present Findings

Parse the JSON output and present findings grouped by severity:

• HIGH — Must be fixed before committing. Identify the flagged file and line, explain what was found (e.g., "AWS secret key on line 14 of config.py"), why it is dangerous, and the specific remediation steps.
• MEDIUM — Warn the user with a clear explanation and ask whether they want to fix it first or proceed.
• LOW — Mention briefly and move on unless the user asks for detail.

For each finding include: file path, line number, finding type, severity, and a plain-English explanation. Consult references/github-security-scanner-hook-reference.md for the full severity table, finding types, and remediation guidance.

If the scan is clean (exit code 0), confirm clearly: "No security issues found in staged files."

Step 4 — Install Pre-Commit Hook (Optional)

Ask the user: "Would you like to install this as a git pre-commit hook so every future commit is scanned automatically?"

If yes:

$PYTHON .github/skills/github-security-scanner-hook/scripts/install-hook.py

The installer:

1. Detects whether .git/hooks/pre-commit already exists — if it does, it offers to Append (default), Overwrite, or Skip rather than silently destroying existing hooks
2. Creates or extends the hook to call scan-staged.py before every commit
3. Makes the hook executable
4. Prints the hook path and a confirmation message

After installation, remind the user:

• The hook lives in .git/hooks/pre-commit — this directory is not tracked by git, so each contributor must run install-hook.py themselves to opt in
• HIGH severity findings block the commit; MEDIUM/LOW prompt a confirmation
• Emergency bypass: git commit --no-verify — see the reference for important caveats

Step 5 — Remediate Issues

For each HIGH or MEDIUM finding the user wants to fix:

1. Open the flagged file at the reported line number
2. For hardcoded credentials: Replace with an environment variable (os.environ["KEY_NAME"]) or a secrets manager call
3. For private keys: Remove the key, revoke it immediately (assume it is compromised), generate a new one, and add the key file pattern to .gitignore
4. For weak cryptography (bandit): Replace insecure functions — e.g., hashlib.md5 → hashlib.sha256, random.token_hex → secrets.token_hex
5. For false positives: Add to the .secrets.baseline allowlist using python -m detect_secrets audit .secrets.baseline
6. Re-stage the fixed files and re-run Step 2 to confirm all HIGH findings are resolved

Additional Resources

• references/github-security-scanner-hook-reference.md — Full severity classification table, detect-secrets plugin mapping, semgrep vulnerability types with remediation examples, .secrets.baseline workflow, false-positive handling, CI/CD integration YAML, --no-verify warning, and secret rotation checklist
• rules/security.yml — Bundled semgrep rules covering SQL injection, command injection, path traversal, XSS, eval/exec, and open redirects across Python, JavaScript/TypeScript, and Go
• scripts/scan-staged.py — Scans staged (or all tracked) files using detect-secrets and semgrep; outputs structured JSON with findings and severity summary
• scripts/install-hook.py — Installs or extends the git pre-commit hook to automate scanning on every future commit

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: nainishshafi
• Source: nainishshafi/developer-productivity-skills
• License: MIT

Install and usage instructions live in the source repository linked above.