The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to audit and process third-party repository content such as README and CHANGELOG files, which are ingestion points for untrusted data. A malicious file could contain instructions designed to manipulate the agent's behavior during the audit. \n
Ingestion points: External repository files are read and parsed by the included Python audit scripts. \n
Boundary markers: The workflow lacks defined delimiters or explicit instructions for the agent to disregard commands embedded within the documents being analyzed. \n
Capability inventory: The skill enables file system modifications and local script execution. \n
Sanitization: There is no evident sanitization of file content before it is processed or presented to the agent. \n
[DYNAMIC_EXECUTION]: The skill utilizes Python scripts that interact with the system environment. \n
Evidence: scripts/audit_semver.py executes git tag via subprocess.run() to extract versioning metadata from the repository context.

repository-documentation