user-research-cookiy

Pass

Audited by Gen Agent Trust Hub on Jun 1, 2026

Risk Level: SAFE
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses a provided shell script, scripts/cookiy.sh, to interface with the Cookiy AI platform. The script uses standard utilities such as curl and jq for networking and data parsing.
  • [EXTERNAL_DOWNLOADS]: The skill communicates with the vendor's official domains, cookiy.ai and s-api.cookiy.ai, to perform research tasks and retrieve platform updates.
  • [CREDENTIALS_UNSAFE]: To maintain authentication, the skill stores access tokens in a local file at ~/.cookiy/token.txt. The script enforces restricted file permissions for this file.
  • [PROMPT_INJECTION]: The automated research synthesis process ingests external interview transcripts, which represent a surface for potential indirect prompt injection. The skill mitigates this through a structured, multi-phase analytical pipeline that focuses the agent on qualitative coding rather than open-ended instruction following.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 1, 2026, 07:04 AM
Security Audit — agent-trust-hub — user-research-cookiy