performance-analysis

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

• Third-party content exposure detected (medium risk: 0.30). In performance-analysis the runtime repeatedly reads the student-owned Meta-ads “Base” data and IT-Systems note from the vault (including human-entered Leads fields) and then generates a verdict snapshot, so outsider-authored free text can flow into the LLM via those user records.

HIGH W008: Secret detected in skill content (API keys, tokens, passwords).

• Secret detected (high risk: 1.00). I scanned the skill and reference files for literal, high-entropy credentials. The file references/changing-the-base.md contains a hardcoded "Master base token" string (Yt2WbfTeQa0tjQsjMUwlfaEvgXb) (line [15]). This value is not a generic placeholder (like YOUR_API_KEY) or a simple setup password — it is a high-entropy token-like string and therefore a real secret that should be treated as a credential and removed or redacted.