kk-curate

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

• Third-party content exposure detected (medium risk: 0.30). The required workflow ingests outsider-authored free text by reading pending session log markdown files from .ai/kenkeep/_sessions/*.md/.ai/kenkeep/sessions/*.md (frontmatter and ## Transcript/proposals content) and then feeding extracted proposals into the CLI primitives for curation.

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

• Potentially malicious external URL detected (high risk: 1.00). The skill repeatedly runs the external runtime command "npx --yes kenkeep@latest" (and variants like "npx kenkeep@latest"), which fetches and executes the remote kenkeep npm package at runtime and is required for core primitives, so remote code controls the agent's behavior.