The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted codebase content which could contain malicious instructions designed to influence the agent's behavior or findings.\n
Ingestion points: Codebase content is ingested via Glob, Grep, and Read tools as defined in references/explorer-prompt.md and SKILL.md.\n
Boundary markers: The prompts for exploration and synthesis do not utilize delimiters or specific instructions to ignore embedded commands within the analyzed files.\n
Capability inventory: Subagents utilize file system exploration tools (Read, Grep, Glob). Security is enhanced by the skill explicitly setting readonly: true for all subagent tasks in SKILL.md.\n
Sanitization: No sanitization or filtering is performed on the data retrieved from the codebase before it is presented to the synthesis or critique agents.

how