Skip to main content

swarm

Use when fanning out parallel subagents across task slices, races, or best-of attempts, not for live web search.

Source facts

Repository
anntnzrb/agents
Last source activity
October 4, 2026 at 16:52
Detected SKILL.md language
English
Stars
1
Forks
0

Install options

The review-first prompt is selected by default. You can switch to a direct command or download a local copy.

Review the source files

Read SKILL.md and any companion files shown by SkillsMP before deciding whether to install.

File Explorer
3 files

Showing SKILL.md

SKILL.md
Source instructions · Read-only preview
disable-model-invocation
true
name
swarm
description
Use when fanning out parallel subagents across task slices, races, or best-of attempts, not for live web search.
license
AGPL-3.0-or-later
# Swarm Fan out N parallel workers. They may cover separate slices, race the same brief, or mix both. The parent waits, aggregates, and returns one report. Adapted from pstack (Lauren Tan, MIT): cloud workers replaced with local subagents. ## Start Open a todolist with one entry per phase before launching anything. 1. Frame 2. Fan out 3. Aggregate 4. Report ## Phase A: Frame 1. State the done predicate and the artifact or report the swarm must return. 2. Choose the shape. Partition into slices, race N workers on identical briefs, or mix both. For a race or mixed shape, declare `first pass`, `rank all`, or `best-of` before spawning. 3. Set N from the user or derive it from the shape. 4. Give each worker its own writable output when it writes. Use a worktree, branch, or `<temp-dir>/swarm-<slug>/worker-<n>/`. N workers writing to the same path is shared mutable state and fails on arrival. When workers verify or measure commits, each brief names the exact SHAs. A measurement brief also names the method (sample count, what one sample is, order). The worker records both in its result. ## Phase B: Fan out Spawn all N workers in parallel when the harness supports concurrent subagents; otherwise execute the briefs sequentially yourself. Each brief stands alone. Include the goal, scope, exact slice or race arm, how to verify, and what to report. Reports use `PASS`, `ISSUES`, or `BLOCKED` with evidence. A worker that can prove a defect reports `ISSUES` and lists every issue it can prove, not only the first. If a worker drops out, proceed with N-1 and note it. ## Phase C: Aggregate Read the terminal results. Drop a result that does not record the SHAs and method its brief names, and rerun that worker once. After a second miss, record a gap. A gap does not count as a pass. For coverage, every required slice needs a result. For a race, apply the selection rule declared up front. Use first pass, rank all, or best-of. Do not paste raw worker dumps. Keep a compact result table, one-line evidenced issues, and explicit gaps or dropouts. ## Phase D: Report Return one consolidated in-chat report with the table, issue one-liners, gaps or dropouts, and the race rule when used.
View on GitHub