The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to analyze project files using Read, Glob, and Grep to generate task plans. This creates a surface where the agent could potentially ingest instructions hidden within a project's source code or documentation.
Ingestion points: Project files accessed via Read, Glob, and Grep tools.
Boundary markers: None specified in the instructions.
Capability inventory: Limited to file read and search operations; no code execution or network transmission capabilities are granted in the skill configuration.
Sanitization: None specified for external content.
[COMMAND_EXECUTION]: The skill references several Python scripts (e.g., ux_audit.py, security_scan.py) as examples of project-specific tooling. However, the skill does not provide these scripts, nor does it command their execution or download. These are used purely as illustrative examples for different project types.

plan-writing