The Agent Skills Directory

[SAFE]: The skill is composed of deterministic Python scripts using only the standard library. No malicious code, obfuscation, or persistence mechanisms were detected.\n- [COMMAND_EXECUTION]: The skill includes instructions for running its scripts via the command line. These commands are limited to local execution and do not involve dangerous parameters or privilege escalation.\n- [PROMPT_INJECTION]: The skill has an indirect prompt injection surface as it processes external RFP documents.\n
Ingestion points: Untrusted text is read by scripts/rfp_parser.py from files provided in the --input argument.\n
Boundary markers: The data is parsed into structured JSON; no explicit boundary markers or 'ignore' instructions are added to the output content.\n
Capability inventory: Across all scripts (rfp_parser.py, response_drafter.py, winrate_predictor.py), there are no subprocess calls, no usage of eval() or exec(), no file-writing outside of stdout redirection, and no network operations.\n
Sanitization: The skill relies on regular expressions for data extraction and standard JSON serialization, which provides inherent structural sanitization.

rfp-responder