The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes content from external Markdown files in the /home/mat/dev/gtd-pcm/ directory. Maliciously crafted tasks within these files could attempt to influence the agent's behavior when the list is generated and returned to the context.
Ingestion points: Daily coaching pulse files (06-coaching/daily/YYYY-MM-DD.md) and project task files (03-projects/*/01-*.md).
Boundary markers: No delimiters or explicit instructions are provided to the model to ignore embedded instructions within the parsed task data.
Capability inventory: The skill is restricted to Glob and Read tools, providing a strong security boundary as it lacks network, file-write, or execution capabilities.
Sanitization: No sanitization or validation logic is defined for the data extracted from the project files.

focus