grant-writing
Pass
Audited by Gen Agent Trust Hub on May 12, 2026
Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill instructions in Step 0 of SKILL.md direct the agent to parse external Notice of Funding Opportunity (NOFO) or Funding Opportunity Announcement (FOA) documents from user-provided URLs. This creates a surface for indirect prompt injection, as untrusted external content could contain malicious instructions designed to influence agent behavior.
- Ingestion points: NOFO/FOA documents processed via URL in SKILL.md.
- Boundary markers: No specific delimiters or instructions to ignore embedded commands are present.
- Capability inventory: The skill utilizes command-line execution (via uvx) and file system access.
- Sanitization: No explicit sanitization or validation of external content is defined.
- [EXTERNAL_DOWNLOADS]: The Additional Resources section in SKILL.md recommends the use of 'uvx opencite' for literature searches. This command involves downloading and executing the 'opencite' package from the Python Package Index (PyPI) at runtime.
- [COMMAND_EXECUTION]: The skill invokes command-line operations, specifically using 'uvx' to run literature search tools and referencing the execution of 'manuscript:humanizer' to refine the writing style of generated documents.
Audit Metadata