The Agent Skills Directory

[COMMAND_EXECUTION]: The skill facilitates diagnostics by running local Python scripts included in the package (scripts/collect.py) and standard system tools like pandoc and typst. It also executes scripts from sibling skills located in the expected installation path.
[INDIRECT_PROMPT_INJECTION]: The skill has an attack surface for indirect injection as it processes user-authored proposal documents and guidelines to troubleshoot failures.
Ingestion points: Reads proposal markdown and workspace guidelines.md files.
Boundary markers: Implements graded disclosure levels (minimal, structure, full) to control data inclusion.
Capability inventory: Has the ability to write to the local file system and execute diagnostic CLI commands via the agent.
Sanitization: The collection script programmatically redacts emails, matriculation numbers, and author names from the collected data using predefined regular expressions.
[EXTERNAL_DOWNLOADS]: The instructions guide the user to update the skill suite from the vendor's registry and to submit bug reports via a pre-filled GitHub issue URL. These interactions are directed toward the vendor's established infrastructure.

proposal-troubleshoot