RiskLens module - Overview
The RiskLens module (sidebar → RiskLens, route #/risklens) is the studio to open before Code. It assumes your plan, launch, hire, or decision already failed 6 months from now, then works backward through every genuine failure reason to produce a revised plan and a pre-launch checklist.
The method comes from psychologist Gary Klein (Harvard Business Review). The core insight: asking "what could go wrong?" yields cautious answers; saying "this already failed - tell me why" switches reasoning into narrative mode and surfaces far more specific causes.
In Navin, RiskLens sits just above the Code module in the sidebar: explain and stress-test the idea before you start the project.
How it works
- Open RiskLens in the sidebar (above Code).
- Click an action card in one of the three groups - Run, Decisions, Focus (see Actions).
- The chat opens with
/risklensand the card prompt already in the composer (height follows the text). Complete the plan in chat, then send. - The agent gathers missing context if needed, sets the failure frame, generates reasons, deep-dives in parallel, synthesizes, then saves the report files.
Direct usage in the module chat:
/risklens launch a billing SaaS at $29/mo for freelancers, 200 paying users in 6 months
The /risklens command
| Command | /risklens [plan|launch|decision] |
| Lifecycle | Agent workflow (runs a full agent turn) |
| Skills preloaded | risklens, multi-agent-orchestration, task-planner |
| Output | Synthesis (most likely failure, most dangerous failure, hidden assumption, revised plan, checklist) + HTML/MD files in the workspace |
Session flow
| Step | What the agent does |
|---|---|
| Context | Checks the minimum bar: what it is, who it is for, what success looks like. Asks focused questions if a piece is missing. |
| Frame | States explicitly: "in 6 months, this plan has already failed." |
| Raw reasons | Lists every genuine failure reason grounded in plan details (no generic padding). |
| Deep-dives | One subagent per reason, in parallel (spawn): failure story, underlying assumption, early warning signs. |
| Synthesis | Most likely failure, most dangerous failure, hidden assumption, concrete revised plan, 3-5 item checklist. |
| Deliverables | risklens-report-[timestamp].html, risklens-transcript-[timestamp].md, short chat summary. |
Good and bad targets
Good targets
- A product or feature about to be built
- A launch with money or reputation on the line
- A pricing or business-model change
- A hire about to be made
- A strategy or positioning pivot
- A partnership or deal under evaluation
- Any commitment where being wrong is expensive
Bad targets
- Vague ideas with no concrete plan yet (plan first, then risklens)
- Questions with one right answer
- Creative feedback on a draft (that is editing)
- Decisions already made and irreversible
Module scoping
/risklens belongs to the RiskLens module. From Code, the command is hidden from the palette and rejected if typed - open RiskLens in the sidebar, then retry. Other studios (/seo, /campaign, …) stay out of RiskLens.
Recommended sequence
- RiskLens - stress-test the plan, get the revised plan and checklist.
- Documents (
/studio) - if you need to formalize the brief or pitch. - Code (
#/code) - only after that, to implement the revised version.
Tips
- Always give the three context pieces: what, who, success.
- The useful product is the synthesis and the revised plan, not a raw fear list.
- Demand revisions the user can run this week ("$47 pilot with 20 people"), not vague advice.
- The agent must not start coding in this module - that is intentional.