The Agent Skills Directory

[COMMAND_EXECUTION]: Provides detailed CLI usage for the trufflehog utility, including subcommands for scanning filesystems, git repositories, and cloud providers like AWS S3 and Google Cloud Storage.
[COMMAND_EXECUTION]: The skill documents the unsafe flag in custom detector YAML configuration which allows bypassing TLS verification for internal verification endpoints, potentially exposing verification requests to man-in-the-middle risks if used on public networks.
[EXTERNAL_DOWNLOADS]: Documents integration with GitHub Actions using the official trufflesecurity/trufflehog repository for automated CI/CD secret scanning.
[CREDENTIALS_UNSAFE]: References the use of environment variables like $GITHUB_TOKEN to authenticate scans against private organization resources, which is the standard expected usage for this security tool.
[INDIRECT_PROMPT_INJECTION]: The skill describes a tool that ingests data from external sources (repositories, filesystems, cloud storage) to identify secrets.
Ingestion points: Reads content from local/remote git repos, S3 buckets, Docker images, and local filesystems (SKILL.md, references/custom-detectors.md).
Boundary markers: None explicitly defined in the documentation for handling tool output.
Capability inventory: The tool performs network operations to verify credentials against third-party APIs.
Sanitization: None mentioned for the data being scanned; however, the tool identifies patterns and entropy rather than executing instructions.

trufflehog