The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted content from external GitHub repositories, which presents a surface for indirect prompt injection attacks.
Ingestion points: The workflow in SKILL.md uses gh search and gh api to retrieve code snippets and file contents from arbitrary repositories.
Boundary markers: There are no instructions in SKILL.md to use delimiters or specific warnings to prevent the agent from potentially following instructions found within the retrieved code (e.g., in comments).
Capability inventory: The skill utilizes the gh CLI tool, base64 decoding, and a utility named opensrc to read and process files (as seen in SKILL.md).
Sanitization: The skill does not implement sanitization or validation of the fetched content before the agent reads it.
[COMMAND_EXECUTION]: The skill utilizes shell commands, specifically the gh CLI, to search, rank, and retrieve repository content.

gh-search