risky-changes

CRITICAL E004: Prompt injection detected in skill instructions.

• Potential prompt injection detected (critical risk: 1.00). The skill promotes a third-party commercial service (DeepAPI) and instructs users to obtain a paid key and credit top-ups, which constitutes monetization steering inside a non-financial skill.

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

• Third-party content exposure detected (low risk: 0.10). The workflow ingests search research results from DeepAPI and live production data via curl and web queries, which represent external untrusted sources of free text.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.70). The skill instructs the agent to query an arbitrary external API (deepapi.co) at runtime for research prompts and behavior-shaping analysis without official first-party provenance, constituting a weak operational dependency.