skills/m13v/s4l/social-autoposter/Gen Agent Trust Hub

social-autoposter

Pass

Audited by Gen Agent Trust Hub on Jul 9, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONDATA_EXFILTRATIONCREDENTIALS_UNSAFE
Full Analysis
  • [EXTERNAL_DOWNLOADS]: Fetches developer tools and automation libraries from established sources, including the 'uv' Python tool from astral.sh and the 'browser-harness' project from GitHub.
  • [COMMAND_EXECUTION]: Utilizes shell commands and Python subprocesses to manage browser profiles, execute automated interactions, and install persistent background services via launchd (macOS) and systemd (Linux).
  • [DATA_EXFILTRATION]: Transmits an 'X-Installation' telemetry header to the author's API at s4l.ai, containing machine metadata and git configuration identifiers as detailed in the project's PRIVACY.md.
  • [CREDENTIALS_UNSAFE]: Accesses a sensitive environment configuration file in a sibling project's directory (~/fazm/web/.env.local) to retrieve third-party API credentials for transcription functionality.
  • [COMMAND_EXECUTION]: Implements self-healing and anti-tampering mechanisms, such as clearing stale browser locks and enforcing macOS immutable flags (uchg) on critical pipeline scripts to prevent unintended modifications.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 9, 2026, 09:06 PM
Security Audit — agent-trust-hub — social-autoposter