wiki-commands

Pass

Audited by Gen Agent Trust Hub on Aug 5, 2026

Risk Level: SAFE
Full Analysis
  • [PROMPT_INJECTION]: The skill ingests untrusted data from external URLs and APIs (such as ArXiv and RSS feeds) through the hermes wiki ingest and hermes wiki monitor commands. This creates a surface for indirect prompt injection where ingested content could attempt to influence the agent's behavior during page generation or classification. However, this is a core functional requirement of a wiki management tool.
  • [REMOTE_CODE_EXECUTION]: The tool supports custom classifier and processor plugins. It includes a specific security mechanism requiring explicit user authorization via hermes wiki plugins trust before any custom code is executed, preventing the unauthorized execution of potentially malicious scripts.
  • [COMMAND_EXECUTION]: The skill describes and facilitates the execution of various CLI commands via the hermes binary. These commands interact with the local file system for wiki storage and the network for ingesting new knowledge base entries.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 5, 2026, 03:01 AM
Security Audit — agent-trust-hub — wiki-commands