agent-setup

Warn

Audited by Gen Agent Trust Hub on Feb 15, 2026

Risk Level: MEDIUMEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [External Downloads] (MEDIUM): The skill instructs the agent to fetch content from https://fullstackrecipes.com/api/recipes/agent-setup using curl. This domain is not recognized as a trusted source, posing a risk of downloading unverified content.\n- [Indirect Prompt Injection] (MEDIUM): The skill facilitates the ingestion of external 'recipes' to configure agent coding guidelines, creating a vulnerability surface where malicious instructions could be embedded in the fetched configuration.\n
  • Ingestion points: Fetches data via a custom MCP resource URI and an external API endpoint in SKILL.md.\n
  • Boundary markers: Absent; there are no instructions to isolate or treat the external content as untrusted.\n
  • Capability inventory: Configures project-specific patterns and coding guidelines, which directly affects how the agent generates code.\n
  • Sanitization: None detected; the agent is expected to adopt the patterns as provided.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Feb 15, 2026, 09:38 PM