skills/notdp/hive/gang-worker/Gen Agent Trust Hub

gang-worker

Warn

Audited by Gen Agent Trust Hub on May 6, 2026

Risk Level: MEDIUMCOMMAND_EXECUTIONREMOTE_CODE_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill mandates the execution of 'python3 -m pip install -e . --break-system-packages' and 'npx skills add "$PWD" -g --all'. These commands perform persistent system-level changes and bypass standard environment protections (PEP 668) designed to prevent system package corruption.- [REMOTE_CODE_EXECUTION]: The skill retrieves task artifacts from an orchestrator agent ('orch') and proceeds to install and register that code. This creates a functional chain for executing code provided by external sources, which could be malicious.- [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection (Category 8). It ingests untrusted artifact data from the 'orch' agent (Ingestion point: SKILL.md) and interpolates it into the system environment via the mandatory refresh step (Capability: pip/npx execution). There are no boundary markers or sanitization steps documented to prevent malicious instructions within the received artifacts from triggering unintended system actions.
Audit Metadata
Risk Level
MEDIUM
Analyzed
May 6, 2026, 12:27 PM