The Agent Skills Directory

[PROMPT_INJECTION]: The skill contains a vulnerability surface for Indirect Prompt Injection (Category 8) because it processes untrusted data from external repositories.
Ingestion points: Untrusted data enters the agent's context through git clone, git log, and the reading of source files, dependency manifests, and issue descriptions during the analysis phases.
Boundary markers: The instructions lack explicit delimiters or warnings to treat repository content as untrusted, increasing the risk that the agent may follow instructions embedded within the scanned repository's files.
Capability inventory: The skill uses the Bash tool to execute git, gh, and ls commands on the local system.
Sanitization: While the skill performs basic cleaning of URL tracking parameters, it does not sanitize or escape the content of the files it reads before processing them with the LLM.
[EXTERNAL_DOWNLOADS]: The skill is designed to clone external source code repositories from user-provided URLs to temporary directories (/tmp or %TEMP%). This behavior is consistent with the skill's primary purpose as a static analysis tool.

scan-repo