The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes repository content which could potentially contain malicious instructions intended to influence the agent's behavior.
Ingestion Point: The fastcontext tool reads and analyzes files from the user's local repository.
Capability Inventory: The skill utilizes Bash to execute exploration commands.
Boundary Markers: No explicit boundary markers or instructions to ignore embedded content are provided.
Sanitization: There is no evidence of sanitization for the data read from the repository.
[DYNAMIC_CONTEXT_INJECTION]: The skill uses dynamic context injection for a pre-execution environment check.
Evidence: !command -v fastcontext >/dev/null 2>&1 || echo ...`` in SKILL.md. This command checks if the required CLI binary is present on the PATH and displays an installation warning if it is not found.
[COMMAND_EXECUTION]: The skill instructs the agent to invoke a specific CLI tool to perform its primary function.
Evidence: The skill description and usage examples detail the execution of fastcontext with various flags like -q and --citation via the Bash tool.

fastcontext