The Agent Skills Directory

[COMMAND_EXECUTION]: The skill is primarily a documentation resource that provides dozens of examples for executing shell commands using specialized search utilities (rg, fd, sg, rga, tokei, scc). These commands interact with the file system and read file contents.
[EXTERNAL_DOWNLOADS]: The documentation includes instructions for installing tools via common package managers such as apt, brew, cargo, go, and npm. These references point to legitimate, well-known software distributions and official repositories.
[PROMPT_INJECTION]: The skill facilitates the processing of untrusted data from a codebase, which presents an attack surface for indirect prompt injection where malicious instructions hidden in files could influence the agent's behavior.
Ingestion points: Search results from rg, fd, sg, and rga which parse arbitrary file contents and names (e.g., in SKILL.md and references/search-strategies.md).
Boundary markers: Absent; the skill does not provide specific instructions or delimiters to isolate search results or warn the agent to ignore embedded instructions within processed files.
Capability inventory: The skill empowers the agent with capabilities to read, search, and structurally rewrite files (via ast-grep in references/ast-grep-patterns.md), which could be exploited if the agent obeys instructions found in search results.
Sanitization: Absent; there is no guidance for filtering, escaping, or validating file content before it is processed by the agent.

file-search