best-practices-researcher

Pass

Audited by Gen Agent Trust Hub on Mar 3, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection (Category 8) due to the processing of untrusted data from multiple sources.
  • Ingestion points: Uses the Glob tool to read local SKILL.md files across the filesystem (e.g., **/**/SKILL.md) and the Context7 MCP to fetch external documentation from the web and GitHub repositories.
  • Boundary markers: The instructions lack specific delimiters or instructions to the agent to disregard embedded commands or formatting within the ingested text.
  • Capability inventory: The skill has capabilities to read files, perform network-based research, and synthesize data for the user.
  • Sanitization: There is no specified logic for sanitizing or validating the content of the files or web pages before they are processed by the agent.
  • [COMMAND_EXECUTION]: The skill utilizes recursive filesystem traversal via Glob patterns (**/**/SKILL.md) to discover and read the full content of skill files, which may lead to processing unintended files if the agent has broad permissions.
  • [EXTERNAL_DOWNLOADS]: The skill performs extensive network operations to retrieve data from official documentation sites, GitHub, and general web searches to satisfy research requests.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 3, 2026, 11:52 AM