start

A command that runs a structured brainstorming session with four to eight agents, each given a different perspective. It combines their responses into areas of agreement, disagreement, unanswered questions, and a suggested next step.

In plain words
What is it for?
Use it to explore product decisions, plans, risks, or other topics where several perspectives are useful.
Why use it?
It helps expose different viewpoints on a decision without relying on one person's ideas. The structured summary makes the discussion easier to act on.

Command

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add commands/viktorbezdek/skillstack/start
Clone the repo
git clone --depth 1 https://github.com/viktorbezdek/skillstack
Per session 41 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 876 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00041 $0.00876
Opus 5 $0.00020 $0.00438
Sonnet 5 $0.00008 $0.00175
Haiku 4.5 $0.00004 $0.00088

Measured yesterday against content hash aada7fa68b19, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

start scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured yesterday.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

brainstorm-swarm/commands/start.md · 89 lines

How it starts

The opening of the file, as written. The whole thing — 89 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Run a brainstorm-swarm session on the topic provided in $ARGUMENTS.

If $ARGUMENTS is empty, ask the user what they want to brainstorm in one sentence.

Workflow

Follow the brainstorm-swarm plugin's three skills in order:

1. swarm-protocol — choose the persona subset

Load swarm-protocol skill. Pick the persona subset (4-8 of the canonical 12) appropriate to the decision type. Use the default-by-decision-type table from references/invocation-patterns.md.

Show the user the proposed subset with rationale. Get sign-off before spawning. Allow the user to add/remove personas (they may also request a custom persona — invoke the custom-personas skill if so).

2. interview-facilitation — design the prompts

Load interview-facilitation skill. Construct the Phase 1 prompt:

  • One-sentence topic framing
  • 100-300 words of context (background, constraints, the specific question)
  • Length budget: under 400 words per persona

Same prompt body goes to every persona; the persona's subagent type loads their voice.

3. swarm-protocol — spawn in parallel

Single message, multiple Task() calls. Sequential spawning collapses the swarm. All Task() invocations for the chosen subset go in ONE message:

Task(subagent_type="brainstorm-swarm:pm", description="...", prompt="...")
Task(subagent_type="brainstorm-swarm:engineer", description="...", prompt="...")
Task(subagent_type="brainstorm-swarm:designer", description="...", prompt="...")
[etc. for the chosen subset]

4. swarm-synthesis — produce the artifact

Load swarm-synthesis skill. Once all personas have returned, produce the standard artifact:

  • Personas in the swarm (with rationale)
  • Consensus (what 3+ personas agreed on)
  • Dissent (substantive disagreements with underlying questions)
  • Open questions (with "what evidence would resolve" notes)
  • Recommended next move (with alternative)
  • Per-persona contributions (collapsed)

Preserve dissent. Don't force consensus.

5. Optional second round (Phase 2)

Read the full file on GitHub · 89 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. yesterday First seen · 89 lines · 41 tokens per session scan A aada7fa68b19

Subscribe to this mod's changes

start is a command published in the GitHub repository viktorbezdek/skillstack (11 stars, last pushed 2mo ago), licensed MIT. It adds 41 tokens to every session and 876 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.