pjangler-parity-rules

Warn

Audited by Gen Agent Trust Hub on Sep 4, 2026

Risk Level: MEDIUMCOMMAND_EXECUTIONPRIVILEGE_ESCALATIONINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructs the agent to use bun and node to run build, test, and audit tasks (e.g., bun run build, node dist/index.js audit).
  • [PRIVILEGE_ESCALATION]: The instructions describe a fix pattern for systemd unit staleness that involves regenerating and enabling services by running a provisioning script (70-systemd.sh) with elevated flags (FORCE_SYSTEMD=1). System service management typically requires administrative privileges.
  • [INDIRECT_PROMPT_INJECTION]: The skill describes logic that ingests data from local configuration and system files, which could serve as an attack surface if those files are compromised.
  • Ingestion points: Ingests data from agents/hermes/*/role.yaml and systemd unit files (SKILL.md).
  • Boundary markers: None identified in the provided instructions.
  • Capability inventory: Includes file reading (safeReadText), path resolution (realpathSync), and shell script execution (70-systemd.sh, bun, node) across the described tasks.
  • Sanitization: No explicit sanitization or validation of the ingested YAML or unit file content is mentioned.
  • [EXTERNAL_DOWNLOADS]: The skill's provenance metadata references a remote Git repository (git@github.com:delorenj/pjangler.git). This is identified as a vendor-owned resource.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Sep 4, 2026, 06:55 AM
Security Audit — agent-trust-hub — pjangler-parity-rules