Skip to content

Project Background

Production Use and Integrations

Kingfisher is used in production security workflows and integrated into other open source tools. These public references are not an exhaustive list:

  • MongoDB uses Kingfisher in internal security workflows including pre-commit scanning, CI/CD integration, historical code analysis, and cloud and database validation.
  • Prowler uses Kingfisher as an offline secret-scanning engine with optional live validation.
  • MegaLinter distributes Kingfisher as the REPOSITORY_KINGFISHER linter in its standard and security flavors.
  • Osmedeus runs Kingfisher in repository scans and imports its SARIF findings.
  • DefectDojo imports Kingfisher JSON reports for vulnerability management.
  • hk includes a built-in Kingfisher secret-scanning check.
  • Weblate configures Kingfisher as a developer pre-commit hook.

If your organization or project uses Kingfisher and would like to be included, open an issue or pull request.

Lineage and Evolution

Kingfisher started as an internal fork of Nosey Parker, building on its fast Rust and Vectorscan scanning engine. Since then, MongoDB has expanded it into a tool for finding exposed secrets, checking which credentials still work, understanding what they can access, and revoking them where supported.

Kingfisher's additions cover the full workflow:

Area What Kingfisher adds
Live credential validation HTTP, gRPC, and provider-specific checks that distinguish active credentials, inactive credentials, and inconclusive results. Known secrets can also be checked directly with validate.
Blast-radius mapping Provider-specific identity, permission, and resource discovery to help teams understand what an exposed credential can reach. Available during scans and through the standalone blast-radius command.
Credential revocation Explicit revocation of supported credentials, including multi-step provider API workflows and a standalone revoke command.
More source hosts Native enumeration and scanning for GitLab, Azure Repos, Gitea, Bitbucket, and Hugging Face, alongside expanded GitHub support.
Repository content beyond code Scanning of issues, wikis, gists, and snippets, including gist and snippet history where supported.
Cloud storage and containers S3, Google Cloud Storage, Azure Blob Storage, Hugging Face buckets, and Docker/OCI images and image archives.
Collaboration and API platforms Secrets discovery in Jira, Confluence, Slack, Microsoft Teams, and Postman workspaces, collections, and environments.
Content extraction Scanning inside archives, office documents, SQLite databases, and Python bytecode.
Detection and filtering Base64 secret discovery, entropy and checksum checks, and targeted HTML/CSS context verification to reduce noisy matches.
Multi-part credentials Association of nearby keys, secrets, endpoints, and other supporting values so complete credentials can be validated together.
Upstream detector coverage A catalog built from pinned Betterleaks and selected Veles sources, with reproducible generation, archived licenses, and provenance hashes. Kingfisher adds validation, access-map, and revocation capabilities to imported detectors.
Custom rules Support for both Kingfisher YAML and Betterleaks TOML. Kingfisher's rule schema also supports HTTP/gRPC validation, Liquid templates, component dependencies, filters, checksums, and revocation definitions.
CI and developer workflows Baselines and finding fingerprints for tracking known findings, plus staged-file, branch, and commit-range scanning for focused checks.
Repository coverage and audit Per-repository discovery, fetch, and scan status, resolved Git revisions, scan scope, timing, and incremental audit logs.
Visual triage A local browser viewer with finding filters and an interactive blast-radius graph.
Additional report formats Token-efficient TOON for agents and LLMs, BSON, and standalone HTML audit reports, with credential validation and blast-radius context in supported outputs.
Notifications Filtered scan summaries and finding details sent to Slack, Teams, Discord, Mattermost, Google Chat, or generic webhooks.
Rust, Python, and JavaScript/TypeScript SDKs Composable scanning APIs backed by the same core, rules, and scanner crates, with separate validation and revocation. Python and Node call the native engine in process, not through the CLI; detection stays offline. The prepared stable Node SDK release 1.1.0 is versioned independently.
Native Windows support Windows builds alongside Linux and macOS, across x64 and ARM64.

The application has changed underneath these features, too. Kingfisher replaces the SQLite-backed scan datastore with in-memory findings storage and optional temporary disk offload. Its scan pipeline brings discovery, detection, validation, access mapping, and reporting into one run, backed by shared core, rules, and scanner crates, reusable scanner pools, and bounded regex confirmation.

See Architecture for the current implementation and Changelog for release-by-release development.

Roadmap and Contributions

Ongoing work includes broader upstream detector coverage, more scan targets, and deeper safe response capabilities. File a feature request or read CONTRIBUTING.md to propose or implement an improvement.

Security vulnerabilities should be reported according to SECURITY.md. Kingfisher is licensed under the Apache License 2.0.