framing-effect
Pass
Audited by Gen Agent Trust Hub on Aug 22, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill processes user-supplied text during its analytical phases, creating a surface for indirect prompt injection.\n
- Ingestion points: The skill ingests untrusted data from the user in SKILL.md to analyze the verbatim language of a framed decision during both the Coach mode interaction and the formal audit process.\n
- Boundary markers: There are no explicit boundary markers or instructions to the agent to disregard potentially malicious commands embedded in the processed user content.\n
- Capability inventory: The skill frontmatter does not restrict tool access via the allowed-tools field, meaning the agent operates with its default set of capabilities while processing this data.\n
- Sanitization: No sanitization or validation of the ingested user input is performed before it is re-evaluated by the agent.\n- [SAFE]: The skill references documentation and metadata hosted on deciqai.com and github.com/deciqai, which are official resources owned by the skill's author and are used for educational context.
Audit Metadata