The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill reads external data from docs/SUMMARY.md and other repository documents which could contain embedded instructions.
Ingestion points: docs/SUMMARY.md and task-relevant documentation files.
Boundary markers: None identified. The skill does not include instructions to ignore directives found within the documentation.
Capability inventory: The skill is explicitly restricted to asking questions and is prohibited from writing code, modifying files, or producing plans. This significantly limits the impact of any injection.
Sanitization: No sanitization or validation of the documentation content is performed before processing.

ask