The Agent Skills Directory

[SAFE]: A thorough review of the skill's instructions, templates, and reference guides revealed no malicious patterns, behavior overrides, or hidden commands.
[COMMAND_EXECUTION]: The skill includes a Python CLI tool (scripts/product_scorer.py) for RICE and PMF prioritization. The script was audited and found to be safe: it uses only the Python standard library, contains no network or shell execution logic, and enforces strict validation rules on input data, such as requiring specific numeric ranges for impact and confidence scores.
[DATA_EXPOSURE]: No hardcoded credentials, API keys, or sensitive file path references were detected. The skill does not perform any network operations that could facilitate data exfiltration. External links provided in data/sources.json are limited to well-known and reputable industry resources.
[INDIRECT_PROMPT_INJECTION]: The skill's primary vulnerability surface is the ingestion of user-provided feature data. This risk is effectively mitigated by the scoring script's requirement for structured JSON input and its validation of data types and values, preventing the injection of natural language instructions into the prioritization logic.

product-management