rfp-responder
Pass
Audited by Gen Agent Trust Hub on Nov 1, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No malicious patterns detected. The skill provides local utility scripts for processing RFP documents without any external network calls or obfuscated code.
- [COMMAND_EXECUTION]: The skill utilizes local Python scripts (rfp_parser.py, response_drafter.py, winrate_predictor.py) for its core functionality. These scripts use only the Python standard library and perform deterministic text processing and calculation without risk of command injection or unauthorized shell access.
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted RFP text for parsing. While this constitutes a data ingestion surface, the processing is handled by regex-based Python scripts rather than direct LLM parsing of raw text. The scripts lack dangerous tools (e.g., network access, file system modification beyond input reading), effectively isolating the agent from potential prompt injection instructions embedded in RFP documents.
- [DATA_EXFILTRATION]: All data processing occurs locally. There are no network requests, API calls, or tools configured that would allow for the exfiltration of sensitive RFP or deal data.
Audit Metadata