Skip to content
Back to skills

Swarm

ASecurity

Use for research, audits, sweeps and code mapping where completeness matters. Split the territory into non-overlapping chunks, run one read-only agent per chunk in parallel, then reconcile every result against the code, the decisions, and each other.

  • 9 stars
  • 0 votes
  • 0 copies
  • 0 views
  • Added November 10, 2026
ai-agents

Security analysis

A100/100

Pro scans all 3 files and shows the line behind each finding

Scanned November 10, 2026

npx -y skills add greenfield-inc/agent-farm --skill swarm --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Swarm?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Swarm
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/greenfield-inc-swarm/badge)](https://www.skillsdirectory.com/skills/greenfield-inc-swarm)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: swarm
description: Use for research, audits, sweeps and code mapping where completeness matters. Split the territory into non-overlapping chunks, run one read-only agent per chunk in parallel, then reconcile every result against the code, the decisions, and each other.
---

# Swarm

Narrow agents with one chunk each, plus a reconciliation pass, give full coverage fast.

## Split

- Map the territory and cut it into non-overlapping chunks that cover everything. Write the coverage map (chunk → agent) before launching.
- **Use 5 to 15 agents, scaled to the territory.** Every real research or audit task gets at least 5.

## Dispatch

- **Launch every agent in one batch.** Use your read-only explorer or researcher subagents, on the strongest model the depth calls for.
- Write each prompt from [references/agent-prompt.md](references/agent-prompt.md): one chunk, one question, evidence for every claim, and a list of what it searched for and found absent.
- Save findings to `evidence/` as they land, and keep working with the person meanwhile.

## Reconcile

Before reporting, run one pass:

- **Verify** the claims that matter against the source. Mark each verified or reported.
- **List conflicts** between agents, the code, and earlier decisions. Resolve what evidence settles; ask the person about the rest, with your recommendation.
- **List gaps:** thin chunks and unanswered questions, each with a follow-up agent or recorded as unknown.
- **Publish** a self-contained HTML page laid out as in [references/reconciliation-template.md](references/reconciliation-template.md).

## Keep to

- Findings stated as fact carry evidence from an agent or from your check.
- A swarm reads only: code, tickets and external systems stay as they are.

Files in this skill

  • SKILL.md1.7 KB
  • references/agent-prompt.md971 B
  • references/reconciliation-template.md844 B

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

View all in ai-agents →

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.