agent-apprenticeship-ecosystem

Fail

Audited by Gen Agent Trust Hub on Jun 22, 2026

Risk Level: HIGHEXTERNAL_DOWNLOADSDATA_EXFILTRATIONCOMMAND_EXECUTIONREMOTE_CODE_EXECUTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill is configured to pull "experience packs," seed tasks, and execution traces from the Forsy-AI/agent-apprenticeship GitHub repository.
  • [DATA_EXFILTRATION]: The skill provides functionality to "contribute" execution bundles to a remote ecosystem. The --auto-share automatic configuration setting allows for the silent upload of local workspace content, including execution traces, agent rollouts, and generated artifacts, to a remote repository.
  • [COMMAND_EXECUTION]: The tool executes various agent CLIs (such as Codex, Cursor, and Claude Code) via subprocess calls. It also supports custom agent definitions through command templates (e.g., --command-template), which provides a surface for the construction and execution of arbitrary shell commands.
  • [REMOTE_CODE_EXECUTION]: The installation instructions encourage the use of npx and npm install -g for the agent-apprenticeship package, which executes code directly from the public npm registry. Additionally, the apprentice learn replay functionality implies the execution of logic contained within pulled experience packs.
Recommendations
  • AI detected serious security threats
Audit Metadata
Risk Level
HIGH
Analyzed
Jun 22, 2026, 12:35 PM
Security Audit — agent-trust-hub — agent-apprenticeship-ecosystem