strategylisted
Install: claude install-skill buildproven/claude-kit
# /bs:strategy - Multi-Model Strategy Synthesis & Advisory Panel
**Usage**: `/bs:strategy "<question>" [--context <file>] [--artifact <file1,file2>] [--providers <list>] [--mode debate|parallel] [--output memo|scorecard|tasks] [--rubric <csv>] [--decision <text>] [--persist <path>] [--rounds <n>]`
Use for: architecture decisions, business/pricing strategy, debugging when stuck, validating assumptions before major refactors.
## Goal
Run a reusable cross-model collaboration pass through `scripts/ensemble-runner.js`, then synthesize the panel output into a single recommendation.
This command is the front door for the broader collaboration layer. Do not hand-roll provider calls inline when the runner can do the fan-out, persistence, and score aggregation once.
## Step 1: Clarify the Decision
Ultrathink before you run anything.
1. Identify the actual decision to make, not just the broad topic.
2. Choose only the context and artifact files needed to answer it.
3. Decide whether the answer should be:
- `memo` for narrative synthesis
- `scorecard` for comparing options against explicit criteria
- `tasks` for action planning after the direction is already chosen
4. Define a rubric when tradeoffs need explicit scoring.
5. Pick providers intentionally:
- `claude` for synthesis and tradeoffs
- `codex` for technical rigor and implementation risks
- `gemini` for market framing and alternative angles
## Step 2: Build the Runner Invocation
Use the shared runner in