The Agent Skills Directory

[COMMAND_EXECUTION] (LOW): The skill uses the Bash tool to execute bun run lint and grep. This is standard for developer tools but constitutes local execution that relies on the integrity of the local package.json and node_modules.
[INDIRECT_PROMPT_INJECTION] (LOW): Evidence Chain:
Ingestion points: Reads content from eslint configuration files and source code via linting output.
Boundary markers: None present; data is interpolated directly into a brief.
Capability inventory: Uses Bash, Read, Glob, and Grep. Side effects are limited to local analysis and plan generation.
Sanitization: No explicit sanitization of function names or file paths extracted from the code.
Risk: A malicious actor could name a function or file in a way that attempts to influence the agent when the brief is processed by /plan-create, though the risk is localized to the agent's reasoning process.

plan-lower-code-complexity