pp-coffee-goat

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 0.90). The skill includes explicit data-exfiltration capabilities (a --deliver webhook: sink and an env-controlled automatic feedback endpoint) that can send local/user data to arbitrary external URLs, which is a high-risk feature for abuse; no obvious hidden eval/remote-code backdoor or obfuscated payloads are present.

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

• Third-party content exposure detected (high risk: 0.72). The required runtime workflow for this skill invokes coffee-goat-pp-cli with --agent, and several commands (e.g., search, watch run, creator-review, transcript-search) rely on synced/upstream roaster and creator corpora that can include outsider-authored free text (product descriptions, transcripts/excerpts) which the CLI then returns as readable JSON fields into the agent’s LLM context.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The skill's prerequisite/install steps instruct running go install on GitHub modules (e.g. github.com/mvanhorn/printing-press-library/library/food-and-dining/coffee-goat/cmd/coffee-goat-pp-cli@latest and github.com/mvanhorn/printing-press-library/library/food-and-dining/coffee-goat/cmd/coffee-goat-pp-mcp@latest), which fetch remote code at install time that will be executed by the CLI and is required for the skill to operate.