configcat

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

• Third-party content exposure detected (high risk: 0.75). Outsider free text can enter the LLM context via runtime-fetched ConfigCat Management API JSON (e.g., scripts/configcat_api.sh → api_request → curl response body printed), which is then included in the agent’s output/context; this JSON is not authored by the operating user.

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

• Secret detected (high risk: 1.00). The .env contains literal, high-entropy credentials (not placeholders). Specifically:
• CONFIGCAT_BASIC_AUTH_USERNAME is a UUID-like value (08de9898-...).
• CONFIGCAT_BASIC_AUTH_PASSWORD is a random/base64-like secret (kQuGQW6M...=).
• CONFIGCAT_AUTHORIZATION_HEADER_BASE64 and CONFIGCAT_AUTHORIZATION_HEADER contain the Base64 encoding / HTTP Basic header derived from those credentials.

These are usable Management API credentials (Basic Auth) and meet the definition of secrets (high entropy, provide access). They are not placeholder/example strings and should be treated as leaked secrets.