what-if-oracle
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONNO_CODE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill manages a workflow that ingests untrusted user input to generate speculative future scenarios.\n
- Ingestion points: User-provided 'What-If' questions are ingested during the framing phase in SKILL.md.\n
- Boundary markers: The framework includes a mandatory sharpening and confirmation step where the agent presents the interpreted question to the user for validation before generating branches, serving as a functional boundary.\n
- Capability inventory: The skill instructions (SKILL.md) and reference files (references/scenario-templates.md) contain no executable code, subprocess calls, file-writing commands, or network operations.\n
- Sanitization: The refinement process in Phase 1 acts as a logical sanitization step for user-provided data.\n- [NO_CODE]: The skill consists entirely of instructional markdown and templates without any associated scripts, executables, or package dependencies.
Audit Metadata