intent
Pass
Audited by Gen Agent Trust Hub on Jun 18, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [SAFE]: The skill operates as a purely instructional and organizational component for an orchestration loop. It manages the transition between calibration and research phases.
- [NO_CODE]: No executable scripts or binary files are provided with this skill. It consists entirely of a workflow definition and architectural constraints.
- [PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface as it ingests untrusted user goals to generate planning documents used by downstream agents.
- Ingestion points: SKILL.md Step 1 reads "user goal" and "workspace" data as the primary source for plan generation.
- Boundary markers: None identified; user input is processed through the AwN framework without explicit instructions to delimit the input from the agent's core instructions.
- Capability inventory: The workflow involves writing to the file system (intent.md) and generating ledger entries that bind the plan to the execution cycle.
- Sanitization: No explicit sanitization or validation for malicious instructions within the incoming user goal is described beyond classification into the AwN framework classes.
Audit Metadata