The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes a local Python script scripts/check.py to perform deterministic structural checks on the proposal file. This execution is scoped to the provided script and is necessary for the skill's primary function.
[INDIRECT_PROMPT_INJECTION]: The skill processes user-provided markdown files (thesis proposals) that could contain malicious instructions. However, the skill implements strict structural boundaries, uses a deterministic script for the first pass, and provides explicit instructions to the agent to act only as a read-only diagnostic tool, minimizing the risk of instruction override.
Ingestion points: Reads a user-specified markdown file (proposal.md).
Boundary markers: The script uses a specific SHA-256 digest mechanism to detect if the file content was modified during the run, enforcing a read-only mandate.
Capability inventory: The skill uses python3 to run the local check.py script and can read files but is explicitly forbidden from writing or modifying files.
Sanitization: The agent pass is limited to specific checks (typos, grammar, timeframe verification) and is instructed to report findings only in chat without generating new files.

proposal-check