The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the Bash tool to gather factual information from the codebase for context grounding. While a powerful capability, it is restricted to repository inspection for the primary purpose of requirement clarification.
[PROMPT_INJECTION]: The skill processes project-level documentation, which represents an indirect prompt injection surface. 1. Ingestion points: Reads AGENTS.md, CLAUDE.md, README, and other files in docs/ during Phases 0 and 2. 2. Boundary markers: The skill does not explicitly instruct the agent to use delimiters or ignore embedded instructions within these ingested files. 3. Capability inventory: The agent has access to Bash, Read, Grep, and Glob tools and can write to the local filesystem. 4. Sanitization: There is no defined process for filtering or validating the content of the documents being processed.

deep-interview