ai-strategist

Warn

Audited by Snyk on Feb 21, 2026

Risk Level: MEDIUM
Full Analysis

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

  • Third-party content exposure detected (high risk: 0.90). The skill's Phase 2 workflow and agent prompt templates (references/agent-prompts.md and Phase 2 in SKILL.md) explicitly instruct researcher agents to use WebSearch/WebFetch and scrape public sources like GitHub, Product Hunt, Hacker News, Reddit, Twitter/X, developer blogs and forums, and those collected outputs are then read and scored by downstream specialists to drive tool selection and the integration roadmap—clearly exposing the agent to untrusted, user-generated third-party content that can influence actions.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Feb 21, 2026, 09:24 AM