The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses local Python scripts (e.g., scripts/rfp_response_analyzer.py) and short inline Python commands (python -c) to parse JSON output and calculate scores. These operations are intended for automation and do not involve untrusted remote sources or privilege escalation.- [DATA_EXPOSURE_AND_EXFILTRATION]: All data processing is performed locally on assets such as assets/sample_rfp_data.json. No network exfiltration commands (curl, wget), hardcoded credentials, or access to sensitive system directories (e.g., .ssh, .aws) were identified.- [INDIRECT_PROMPT_INJECTION]: The skill ingests data from RFP and competitive JSON files. Ingestion points: Data enters via assets/sample_rfp_data.json and competitive_data.json. Boundary markers: No explicit markers are used to delimit data from instructions. Capability inventory: The skill executes local Python scripts and performs file writes. Sanitization: No explicit sanitization is documented for the JSON content. However, the use of structured data processing via scripts minimizes the risk of the agent misinterpreting data as instructions.

sales-engineer