make-proposals

Generate two solution proposals grounded in existing research and a feature or project request.

Behavior

• Require a research document produced by /research-codebase (either a local file path or a Linear document).
• Produce two distinct solution approaches, each tied back to the research findings.
• Highlight trade-offs, impacted systems, validation steps, and open questions per proposal.

Steps

1. Intake
   • Capture the user's request. If a Linear issue or project is referenced and Linear MCP tools are available, read it.
   • Read the research document (local file or Linear document).
2. Parse research
   • Extract constraints, relevant modules, dependencies, data flows, and prior decisions.
3. Synthesize solution space
   • Derive candidate approaches grounded in the research findings.
   • For each approach, note primary changes, affected code paths, required migrations/config updates, and rollout considerations.
   • Keep the list to two distinct proposals, ordered from most to least aligned with constraints.
4. Validation planning
   • Identify verification strategies (tests, experiments, observability) necessary to prove each approach.
   • Surface critical unknowns or prerequisite research.
5. Save proposals:
   • If Linear MCP tools are available, save as a Linear document.
   • Also save locally for reference.
6. Present the proposals to the user for a decision.

Output template

## Solution Proposals

Context:
- Request: <short restatement of the ask>
- Research Source: <document or filename and key sections used>

Proposal 1 — <title>
- Overview: <2-3 sentences>
- Key Changes: <components/modules>
- Trade-offs: <risks vs benefits>
- Validation: <tests/experiments/metrics>
- Open Questions: <gaps or follow-ups>

Proposal 2 — <title>
...

Notes

• Reference code and sections using path/to/file.py:lines when citing specifics.
• When the research does not directly address the request, call out the gaps and suggest additional research before proposing solutions.
• Stay concise; favor clarity over exhaustive detail.