squad-init

Pass

Audited by Gen Agent Trust Hub on Jul 2, 2026

Risk Level: SAFEDATA_EXFILTRATIONCOMMAND_EXECUTIONCREDENTIALS_UNSAFE
Full Analysis
  • [DATA_EXFILTRATION]: The skill extracts the project's Git remote origin URL and transmits it to a remote API (squad-api-285415501393.asia-south1.run.app) during registration.
  • [DATA_EXFILTRATION]: The skill reads the first few hundred characters of the CLAUDE.md file to infer the project's purpose and technology stack, which are then sent to the remote API as part of the project metadata.
  • [CREDENTIALS_UNSAFE]: The skill accesses a sensitive credential file at ~/.squad/auth to verify the presence of a SQUAD_AUTH_TOKEN required for API communication.
  • [COMMAND_EXECUTION]: The skill executes various shell commands, including git remote, grep, sed, and jq, to parse project information and construct JSON payloads safely for the API request.
  • [COMMAND_EXECUTION]: The skill invokes an api tool to perform a POST request to the vendor's remote endpoint.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 2, 2026, 11:34 AM
Security Audit — agent-trust-hub — squad-init