notion-cli
Warn
Audited by Gen Agent Trust Hub on Nov 7, 2026
Risk Level: MEDIUMPROMPT_INJECTIONEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill includes instructions that direct the agent to modify the skill's own file (SKILL.md) automatically. This mechanism allows the agent to override its original instructions based on external triggers or perceived drift. Evidence: 'Self-Evolving Skill' header and 'Post-Execution Reflection' section in SKILL.md.
- [EXTERNAL_DOWNLOADS]: The skill recommends the installation of a binary from a third-party GitHub repository that is not a recognized technology provider or well-known service. Evidence: 'brew install 4ier/tap/notion-cli' in SKILL.md.
- [REMOTE_CODE_EXECUTION]: The skill encourages downloading and running code from an external, untrusted source via a package manager. Evidence: Installation and execution of the 'notion' binary from the '4ier/notion-cli' repository.
- [INDIRECT_PROMPT_INJECTION]: The skill processes potentially untrusted content from the Notion API and has access to powerful tools. Ingestion points: Search, page view, database query, and block commands in SKILL.md. Boundary markers: None. Capability inventory: Bash tool access and filesystem write capabilities. Sanitization: None.
Audit Metadata