hermes-labyrinth-observability

Warn

Audited by Gen Agent Trust Hub on May 17, 2026

Risk Level: MEDIUMEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill clones an external repository from "https://github.com/stainlu/hermes-labyrinth.git" during the installation and Docker setup process.
  • [REMOTE_CODE_EXECUTION]: After downloading the repository, the instructions require running build and test scripts ("npm run build", "python3 scripts/test-plugin-api.py"), which executes code from the untrusted external source.
  • [COMMAND_EXECUTION]: The skill executes various shell commands for directory management ("mkdir -p ~/.hermes/plugins"), cleanup ("rm -rf"), and local API interaction via "curl".
  • [DATA_EXFILTRATION]: The plugin is designed to read sensitive local data from "/.hermes/state.db", "/.hermes/skills/", and "~/.hermes/cron/". While it uses local API endpoints and claims to redact secrets, the potential for data exposure exists if the source repository is compromised.
  • [SAFE]: The skill documentation explicitly mentions and implements a "fail-closed" secret redaction mechanism using the "hermes.redactor" library, which is a defensive best practice.
Audit Metadata
Risk Level
MEDIUM
Analyzed
May 17, 2026, 08:47 PM
Security Audit — agent-trust-hub — hermes-labyrinth-observability