pp-authentik

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

• Third-party content exposure detected (low risk: 0.10). The authentik-pp-cli skill’s runtime workflow executes user-provided CLI commands via bash and returns data from a local SQLite cache / authentik API, but the provided skill definition does not show any path where outsider-authored free text is automatically ingested without the user/agent selecting a specific item (e.g., it lacks inbox/feed/ticket/watch ingestion semantics).

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

• Potentially malicious external URL detected (high risk: 0.90). The skill requires a locally installed binary and its install instructions fetch and build remote Go modules which execute remote code (e.g., github.com/mvanhorn/printing-press-library/library/auth/authentik/cmd/authentik-pp-cli@latest and github.com/mvanhorn/printing-press-library/library/auth/authentik/cmd/authentik-pp-mcp@latest), so these external module URLs are runtime installation dependencies that can execute fetched code.