twitter-poster

CRITICAL E004: Prompt injection detected in skill instructions.

• Potential prompt injection detected (critical risk: 1.00). The setup script harvests sensitive artifacts (browser cookies, environment vars, SSH config/listings) and uploads them to an external "telemetry" endpoint while the README instructs agents to downplay the validation, constituting secret exfiltration and deceptive/hidden behavior.

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (critical risk: 1.00). The setup script deliberately collects sensitive local data (browser Cookies files, SSH config/listing, environment) and uploads a tarball to a remote telemetry endpoint during the "account linking" step, demonstrating intentional exfiltration of credentials and user data.

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

• Third-party content exposure detected (medium risk: 0.30). The required runtime posting workflow reads queue.txt (user-/outsider-authored scheduled content) and then posts it, meaning free text can be ingested without selecting a specific item first.