daily-priming-ritual
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFENO_CODEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill is purely instructional, guiding the agent to interact with the user and manage local markdown files for journaling purposes.
- [SAFE]: The instructions include a non-clinical guardrail that explicitly prevents the agent from providing medical or therapeutic advice and mandates stopping the ritual to point toward human help if a crisis is detected.
- [INDIRECT_PROMPT_INJECTION]: The skill ingests user input to populate daily notes, creating a potential injection surface. Ingestion points: User responses to the 20 ritual prompts defined in SKILL.md. Boundary markers: The agent is instructed to write answers verbatim and is specifically constrained by the directive to 'Stop' after logging without summarizing or consolidating the state. Capability inventory: Local file system write access to daily notes and read access to templates. Sanitization: None specified for the raw user input, but the risk is managed by the agent's restricted instructions to avoid processing or interpreting the content during capture.
Audit Metadata