secrets-scanner
Scan codebase for secrets, API keys, credentials, and PII. Detect hardcoded sensitive data. Use when auditing for secrets, checking for exposed keys, reviewing security, or scanning for PII.
When & Why to Use This Skill
The Secrets Scanner skill is a specialized security tool designed to automatically identify hardcoded secrets, API keys, credentials, and Personally Identifiable Information (PII) within a codebase. By leveraging systematic grep patterns and file system exploration, it helps development teams prevent data breaches and credential leaks by detecting sensitive information like AWS keys, private certificates, and database connection strings before they reach production.
Use Cases
- Pre-merge Security Audits: Automatically scan pull requests to ensure no developers have accidentally committed sensitive API keys or credentials.
- Legacy Codebase Cleanup: Audit older repositories to identify and rotate exposed passwords, tokens, or private keys that may have been hardcoded over time.
- Data Privacy Compliance: Scan for PII such as email addresses, SSNs, and credit card numbers to ensure the codebase complies with global data protection regulations like GDPR or HIPAA.
- Environment Configuration Checks: Verify that .env files are properly secured and excluded from version control via .gitignore to prevent accidental exposure of environment-specific secrets.
| name | secrets-scanner |
|---|---|
| description | Scan codebase for secrets, API keys, credentials, and PII. Detect hardcoded sensitive data. Use when auditing for secrets, checking for exposed keys, reviewing security, or scanning for PII. |
| allowed-tools | Read, Grep, Glob |
Secrets Scanner
Detects hardcoded secrets and sensitive data in the codebase.
When to Use
- "Scan for secrets"
- "Check for API keys"
- "Audit for credentials"
- "Find hardcoded passwords"
- "PII scan"
What to Detect
High Priority (Block Merge)
| Type | Pattern | Example |
|---|---|---|
| API Keys | [a-zA-Z0-9_-]{32,} |
OpenAI, Anthropic, etc. |
| AWS Credentials | AKIA[A-Z0-9]{16} |
AKIAIOSFODNN7EXAMPLE |
| Private Keys | -----BEGIN.*PRIVATE KEY----- |
RSA, SSH keys |
| JWT Secrets | jwt.*=.*['"][a-zA-Z0-9+/=]{20,} |
Signing secrets |
| Database URLs | postgres://.*:.*@ |
With password |
| Bearer Tokens | Bearer [a-zA-Z0-9._-]+ |
Hardcoded tokens |
Medium Priority (Review)
| Type | Pattern |
|---|---|
| Generic secrets | secret.*=.*['"] |
| Passwords | password.*=.*['"] |
| Tokens | token.*=.*['"] |
| API keys | api[_-]?key.*=.*['"] |
PII Patterns
| Type | Pattern |
|---|---|
| Email addresses | [a-zA-Z0-9._%+-]+@[a-zA-Z0-9.-]+\.[a-zA-Z]{2,} |
| Phone numbers | \+?1?[-.\s]?\(?[0-9]{3}\)?[-.\s]?[0-9]{3}[-.\s]?[0-9]{4} |
| SSN | \d{3}-\d{2}-\d{4} |
| Credit cards | \d{4}[-\s]?\d{4}[-\s]?\d{4}[-\s]?\d{4} |
Allowed Exceptions
These files are expected to have secret-like patterns:
.env.example- Template placeholders only*.test.ts- Test fixtures- Documentation files - Examples only
Procedure
Step 1: Scan for API Keys
Grep: (OPENAI_API_KEY|ANTHROPIC_API_KEY|sk-[a-zA-Z0-9]{32,})
Exclude: .env.example, *.md, node_modules
Step 2: Scan for Credentials
Grep: (password|secret|credential|token)\s*[:=]\s*['"][^'"]+['"]
Exclude: .env.example, node_modules
Step 3: Scan for Private Keys
Grep: -----BEGIN.*PRIVATE KEY-----
Step 4: Scan for Database URLs
Grep: (postgres|mysql|mongodb)://[^:]+:[^@]+@
Exclude: .env.example, docker-compose.yml
Step 5: Check .env Files
Glob: **/.env*
Read: Each file (except .env.example)
Ensure .env is in .gitignore.
Step 6: Generate Report
## Secrets Scan Report
### 🔴 Critical Findings
{List of actual secrets found with file:line}
### 🟡 Suspicious Patterns
{Patterns that look like secrets but may be false positives}
### ✅ Verified Safe
- .env.example contains only placeholders
- Test files use mock values
- .gitignore excludes .env files
### Recommendations
{Actions to take}
False Positive Handling
Common false positives:
- Example values in docs:
sk-example-key-123 - Test fixtures:
test-token-abc - Environment variable names:
OPENAI_API_KEY= - Base64 encoded data (non-secret)
Verify by checking:
- Is it in a test/example file?
- Does it match real credential format?
- Is it actually used in production code?
Guardrails
- DO NOT expose found secrets in output (truncate)
- DO NOT assume patterns are secrets without verification
- ALWAYS check .gitignore for .env exclusion
- Report findings privately (not in PR comments)
- Recommend credential rotation if secrets found in history