detect-secrets

TLDR

Scan a directory for secrets

$ detect-secrets scan [path/to/directory]

Scan and create baseline file

$ detect-secrets scan > [.secrets.baseline]

Audit baseline file

$ detect-secrets audit [.secrets.baseline]

Scan specific files

$ detect-secrets scan [file1] [file2]

Scan with all plugins

$ detect-secrets scan --all-files [path]

Update existing baseline

$ detect-secrets scan --baseline [.secrets.baseline]

Run as pre-commit hook

$ detect-secrets-hook --baseline [.secrets.baseline]

SYNOPSIS

detect-secrets command [options] [path]

detect-secrets is a tool for detecting secrets (passwords, API keys, tokens) accidentally committed to code repositories. It scans files for patterns matching known secret formats and flags potential exposures.
The tool uses multiple detection plugins: high-entropy strings, keyword patterns (password, api_key), and format-specific detectors (AWS keys, private keys, etc.). A baseline file tracks known secrets to avoid repeated alerts on intentional entries.
The audit workflow allows reviewing detected secrets interactively, marking them as true positives, false positives, or skipped. Pre-commit hook integration prevents new secrets from being committed.

PARAMETERS

scan

Scan files for potential secrets.

audit

Interactively audit detected secrets.

--baseline file

Use existing baseline for comparison.

--all-files

Scan all files, not just tracked by git.

--exclude-files regex

Regex pattern for files to exclude.

--exclude-lines regex

Regex pattern for lines to exclude.

--exclude-secrets regex

Regex pattern for secret values to exclude.

--list-all-plugins

List available detection plugins.

--no-verify

Disable verification of detected secrets.

-f, --force-use-all-plugins

Enable all detection plugins.

PLUGINS

ArtifactoryDetector: Artifactory tokens
AWSKeyDetector: AWS access keys
Base64HighEntropyString: High-entropy base64
BasicAuthDetector: Basic auth credentials
HexHighEntropyString: High-entropy hex strings
JwtTokenDetector: JWT tokens
KeywordDetector: Secret keywords
PrivateKeyDetector: Private key headers
SlackDetector: Slack tokens

CAVEATS

High-entropy detectors may produce false positives on random strings, UUIDs, or hashes. Baseline must be maintained as code changes. Some detection relies on patterns that may not catch obfuscated secrets. Does not scan git history by default.

HISTORY

detect-secrets was developed at Yelp and open-sourced in 2018. It was created to prevent secret leakage in their large codebase. The plugin architecture allows extending detection capabilities. The tool has been adopted widely as part of secure development practices and CI/CD pipelines.