detect-diminishing-returns
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSNO_CODE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructions involve analyzing external research or analysis content, creating a surface where instructions embedded in that data could influence agent behavior.
- Ingestion points: Research data is ingested via the user command in README.md.
- Boundary markers: No specific delimiters or safety warnings are used to separate external data from instructions.
- Capability inventory: The skill does not provide any tool-use capabilities, shell access, or executable scripts.
- Sanitization: No input filtering or sanitization is specified for the processed research data.
- [EXTERNAL_DOWNLOADS]: The README.md provides installation instructions that download content from the author's own GitHub repository using npx.
- [NO_CODE]: The skill consists entirely of natural language instructions in markdown format and does not include any scripts, binaries, or automated tool configurations.
Audit Metadata