The Agent Skills Directory

[SAFE]: The skill acts as a security control by gating sensitive agent actions with human approval. The logic enforces a full stop if consent is not granted or the request expires.
[EXTERNAL_DOWNLOADS]: The skill communicates exclusively with the vendor-owned API at https://api.consent.app. This network activity is documented and essential for the skill's functionality.
[COMMAND_EXECUTION]: All scripts use Python's standard library (urllib, json, argparse) with no external dependencies. There is no evidence of unsafe shell execution, subprocess spawning with untrusted input, or remote code execution.
[CREDENTIALS_UNSAFE]: The skill handles a user-managed API key (CONSENT_API_KEY) using environment variables or CLI arguments. It includes explicit instructions against persisting the key to disk and provides guidance for the user to revoke keys through their mobile app, aligning with security best practices.

consent-app