The Agent Skills Directory

[SAFE]: The skill instructions and associated reference documents (such as reasoning-engine.md and product-playbooks.md) provide purely textual and logical guidance for product management scenarios without invoking dangerous system capabilities.
[COMMAND_EXECUTION]: The skill includes a Python script scripts/quality_gate.py intended for developers to validate response samples. This script uses safe, standard library modules (re, pathlib, sys) to perform regex matches against text files provided as command-line arguments. It does not perform network operations, access sensitive system paths, or execute arbitrary code from untrusted sources.
[PROMPT_INJECTION]: The SKILL.md and references/methodology-basis.md files contain strict persona instructions to ensure the agent remains in character and does not disclose its underlying logical framework. These are standard formatting and character-consistency constraints rather than malicious attempts to bypass AI safety filters.
[DATA_EXFILTRATION]: No patterns indicative of data exfiltration were found. The skill does not contain hardcoded credentials, and there are no instructions to transmit user data to external servers.

product-decision-agent