skills/nousresearch/hermes-agent/qmd/Gen Agent Trust Hub

qmd

Pass

Audited by Gen Agent Trust Hub on Sep 28, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONPERSISTENCEPRIVILEGE_ESCALATION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill instructs the user to install the qmd package from the NPM registry and fetches Node.js installation scripts from the well-known NodeSource repository.
  • [COMMAND_EXECUTION]: The skill uses terminal commands to manage the document index, perform searches, and configure environment settings for the local search engine.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied text documents, markdown notes, and transcripts, creating a surface for indirect instructions to enter the agent's context during retrieval.
  • Ingestion points: Local files and directories added to the search index via qmd collection add.
  • Boundary markers: No specific delimiters or warnings for embedded instructions are defined in the skill body.
  • Capability inventory: The skill allows the agent to execute terminal commands and use MCP tools for document retrieval and status checking.
  • Sanitization: No explicit sanitization or filtering of document content is described.
  • [PRIVILEGE_ESCALATION]: Setup instructions include the use of sudo for system-level package management, specifically for installing Node.js and SQLite extensions.
  • [PERSISTENCE]: The skill provides configuration templates for systemd (Linux) and launchd (macOS) to maintain a persistent background daemon for the MCP (Model Context Protocol) server.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 28, 2026, 12:57 AM