The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes a comprehensive set of git and shell commands to perform deep audits of repository contents and commit history.
[EXTERNAL_DOWNLOADS]: The skill references and provides installation templates for well-known security utilities like gitleaks and trufflehog from reputable GitHub sources.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted repository content through shell commands such as git grep and git log in references/scan-patterns.md (Ingestion points). While documentation templates warn users about secret exposure, boundary markers or 'ignore' instructions are not applied during the automated scanning process (Boundary markers). The agent possesses shell execution capabilities as defined in SKILL.md to perform its audit tasks (Capability inventory), and the skill does not implement specific sanitization of the scanned repository content (Sanitization).
[SAFE]: The skill demonstrates security best practices for public repository management and remediation, with no evidence of malicious behavior or obfuscation.

repo-sentinel