competitive-intelligence

Pass

Audited by Gen Agent Trust Hub on Jul 31, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill is designed to ingest and synthesize data from external, untrusted sources such as competitor websites, changelogs, job postings, and community channels. This creates a potential surface for indirect prompt injection where an adversary could place malicious instructions in publicly monitored content to influence the agent's analysis.
  • Ingestion points: External sources are explicitly listed in SKILL.md under 'Sources available', including websites, docs, press, and community channels.
  • Boundary markers: The instructions do not define technical delimiters for external content but do include behavioral guidance to 'separate observed signals from inference' and 'label evidence vs inference' to mitigate accidental obedience.
  • Capability inventory: No dangerous tools (shell execution, file writing, or network POST requests) are utilized by this skill.
  • Sanitization: No specific filtering or sanitization logic is provided for the ingested data.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 31, 2026, 11:46 AM
Security Audit — agent-trust-hub — competitive-intelligence