devrel-competitor-analysis

Pass

Audited by Gen Agent Trust Hub on Sep 13, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill creates a surface for indirect prompt injection by ingesting untrusted external competitor data, but includes explicit guidance to handle it safely.
  • Ingestion points: The skill fetches competitor documentation, marketing pages, blogs, and public repositories for analysis (SKILL.md, signal-collection.md).
  • Boundary markers: Explicit instructions to treat fetched pages as untrusted input and instructions within them as data rather than commands (Collection discipline).
  • Capability inventory: Agent capabilities for web browsing and forge API access.
  • Sanitization: Instructions to identify and report embedded instructions as findings.
  • [SAFE]: The audit found no evidence of credential theft, malicious persistence, privilege escalation, or obfuscated code.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 13, 2026, 09:14 PM
Security Audit — agent-trust-hub — devrel-competitor-analysis