Skills
Navigate
skills/reason-machines/ai-agent-skills/500-ai-agents-projects-catalog/Gen Agent Trust Hub

500-ai-agents-projects-catalog

Pass

Audited by Gen Agent Trust Hub on Oct 8, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill provides implementation patterns that ingest markdown content from an external third-party repository.\n
  • Ingestion points: The functions find_use_cases_by_industry and get_framework_examples in SKILL.md use requests.get to retrieve data from a remote URL.\n
  • Boundary markers: The provided code snippets lack boundary markers or explicit instructions to the agent to disregard commands embedded within the fetched text.\n
  • Capability inventory: The skill provides examples of using this external data to drive agents using LLMs (OpenAI) and frameworks like CrewAI and AutoGen.\n
  • Sanitization: The patterns demonstrate parsing raw markdown content without showing validation or sanitization steps to filter out potentially malicious prompt content.\n- [EXTERNAL_DOWNLOADS]: The skill fetches data from an external repository to serve as its primary information source.\n
  • Evidence: requests.get("https://raw.githubusercontent.com/ashishpatel26/500-AI-Agents-Projects/main/README.md") in SKILL.md.\n
  • Context: While GitHub is a well-known service, the skill establishs a runtime dependency on the content and integrity of a third-party source not controlled by the vendor.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 8, 2026, 11:38 PM