grant-proposal-assistant
Pass
Audited by Gen Agent Trust Hub on Sep 23, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied grant proposal files, creating a vulnerability surface for indirect prompt injection.\n
- Ingestion points: The "scripts/main.py" script reads external content from file paths provided via the "--input" CLI argument.\n
- Boundary markers: The script does not utilize delimiters or specific instructions to isolate untrusted data from the agent's reasoning logic.\n
- Capability inventory: The skill has the capability to execute a local Python script ("scripts/main.py") which performs file read/write operations on the local system.\n
- Sanitization: Input content is processed (word count and keyword scanning) and then presented back to the agent without escaping or filtering of the source text.\n- [COMMAND_EXECUTION]: The skill instructions direct the agent to execute shell commands to compile and run the provided "scripts/main.py" file. This is the primary intended mechanism for performing proposal reviews and template generation, and it relies on standard Python library modules.
Audit Metadata