ops-linkedin

Pass

Audited by Gen Agent Trust Hub on Aug 26, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [COMMAND_EXECUTION]: The lead_resolver.py module invokes a sibling skill script (brave_search.py) via subprocess.run to perform public web searches for profile hypothesis generation. The execution uses a hardcoded path within the user's home directory and passes search terms as command-line arguments.
  • [EXTERNAL_DOWNLOADS]: The skill utilizes the uv package manager and run.sh to download and manage Python dependencies from standard package registries as defined in the pyproject.toml and lock files.
  • [SAFE]: The skill implements a robust security architecture including an outbound roundtable gate, mandatory human attestation for all actions, and explicit guardrails that prohibit cookies, scraping, and automated social actions on the LinkedIn platform.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 26, 2026, 06:00 PM
Security Audit — agent-trust-hub — ops-linkedin