mktg-x

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 0.90). The code intentionally includes local-browser credential extraction and decryption (Keychain/DPAPI/keyring bridges, cookie DB reading, and inline cookie accepts) and then uses those session cookies to make authenticated requests to X — functionality that is legitimate for the skill but is a high-risk capability for credential theft or exfiltration if misused.

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

• Third-party content exposure detected (high risk: 0.85). Yes—mktg-x fetches outsider-authored tweet/thread/article text from X’s web GraphQL API at runtime and returns it as JSON fields like tweet_text/thread_unroll, which the agent then places into the LLM’s grounding memo (“Copy / Prose”).