clarify

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill explicitly instructs the agent to "record the original requirement verbatim" and passes the requirement text into shell commands (e.g., as the task argument), which forces any secrets included by the user to be captured and emitted verbatim in outputs and CLI calls, creating a direct exfiltration risk.

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 "Web Researcher" sub-agent (in SKILL.md) and the Research Guide (references/research-guide.md, Section 2) explicitly require using WebSearch/WebFetch to retrieve web/best-practice sources and have the agent read and use those findings as evidence for decisions (e.g., Tier 2 classification and expert prompts), which exposes it to untrusted public third-party content that could carry indirect prompt injection.