idea-superpowers-suite
Pass
Audited by Gen Agent Trust Hub on May 13, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill implements a specific override phrase 'GREENLIGHT NEXT STAGE' that allows the user to force the agent to bypass its internal readiness gates and move to the next stage of the workflow even if product or technical gaps remain.
- [PROMPT_INJECTION]: The workflow generates a final 'build handoff' artifact (Stage 7) which contains a 'Prompt for Build Agent' section. This section interpolates user-provided requirements and product vision directly into instructions for a downstream AI agent without sanitization or boundary markers, creating a surface for indirect prompt injection.
- Ingestion points: User-provided text captured during the 'Capture' and 'Interview' stages.
- Boundary markers: Absent; user input is placed directly into the Markdown handoff template.
- Capability inventory: The skill writes multiple Markdown files to the local filesystem across the workflow stages.
- Sanitization: No validation or filtering is applied to user-supplied content before it is included in the final agent prompt.
Audit Metadata