queue

A lock-free shared work queue that MANY workers (terminals/sessions) can drain at the same time without double-processing. Each item is a file claimed by an atomic rename, so concurrent claims are safe with no locks. Supports dynamic enqueue (`add`), a `--stop-file`/budget pause, and `reclaim` of a dead worker's stale claims. USE when work should be pulled by **multiple concurrent workers from one queue** ("process this queue from 3 terminals", "a shared work queue", "let several agents drain these in parallel safely"), or when items arrive over time. For a single orchestrator processing a fixed list, use /agentflow:foreach (its `--shard k/N` splits a list across terminals by static partition — queue is the dynamic, pull model instead).

Ir para a instalação

Repositório: AleSaiani/agentflow
Última atividade na origem: 25 de maio de 2026 às 18:30
Idioma detectado do SKILL.md: inglês
Estrelas: 5
Forks: 0

Opções de instalação

Por padrão, está selecionado o prompt que primeiro revisa a origem. Você pode mudar para um comando direto ou baixar uma cópia local.

Instalar com Codex ou Claude Copie este prompt, cole no Codex, Claude ou outro assistente e deixe que ele revise a página da skill e instale para você.

Um comando direto ignora o prompt de revisão. Verifique a origem antes de executá-lo.

npx skills add https://github.com/AleSaiani/agentflow --skill queue

O comando permanece em uma só linha. Role horizontalmente para revisá-lo antes de copiar.

Prefere uma cópia local? Baixe os arquivos disponíveis atualmente no SkillsMP.

Revise os arquivos de origem

Leia o SKILL.md e os arquivos complementares exibidos pelo SkillsMP antes de decidir se vai instalar.

SKILL.md

Instruções da origem · Visualização somente leitura

Voltar à pesquisa Voltar à instalação

foreach

Apply ONE operation to EVERY item of a list — the **map** (N→N). The operation is a prompt; items come from a JSON list, a markdown checklist, or another run's output. Runs in parallel subagents or inline in the main thread, with persistent state and cross-turn auto-continue. USE on any "do the same thing to each of these" intent — formal OR casual: "for each X…", "per ogni…", "go through these one by one", "do every item in TODO.md", "review each of these files", "handle all of them", a glob, a folder. Trigger on the INTENT, then judge by count (the count gate, Step 2.5): - 1–2 items → just do it inline, no machinery; - a borderline handful (~3–10) → ask whether to use the durable/parallel mechanism or handle inline; - genuinely many, or heavy/independent items → use foreach. DON'T use when the list must first be generated from a spec (→ /agentflow:enumerate, then foreach), or you need one combined output (→ /agentflow:reduce). For the prebuilt "review every file → digest", use /agentflow:audit. Explicit

2026-05-27

run-workflow

Run a declarative workflow-file end to end: init a `/agentflow:pipe` from a JSON `WorkflowSpec`, then drive it to completion — auto-running every bash/json/deterministic stage and dispatching agents only where genuinely needed. Use `--dry-run` to preview the resolved plan without executing anything. The one-command way to execute a reusable, versioned workflow. USE when the user points at a workflow file ("run this workflow", "run workflows/audit/WORKFLOW.md", "run my pipeline") or wants to run a saved `WorkflowSpec`. Workflows are normally self-contained folders (`workflows/<name>/WORKFLOW.md` + sibling scripts), but a bare `.json` path also works.

2026-05-27

audit

Deep code-audit recipe: discover files in a target → review each (LLM, cached) → partition by component → executive digest with hotspots and recurring patterns. A layer-3 recipe over /agentflow:pipe (the 6-stage pipeline and caching are detailed in the body). USE when the user asks to "review / audit / find bugs across" a folder, repo, or glob and wants a persisted, structured report — and the target is large enough (≥ ~10 files) that per-file parallelism plus a digest beats reading inline. DON'T use for a few named files (read them inline), exploratory questions ("what does this do?"), or generic per-item work that isn't a code review (→ /agentflow:foreach). Not a `/`-menu command — it's the shipped `workflows/audit/` recipe: reach it by asking ("audit src for bugs"), or run it directly with `/agentflow:run-workflow workflows/audit/WORKFLOW.md`.

2026-05-26

Ver todas deste repositório

name	queue
user-invocable	false
description	A lock-free shared work queue that MANY workers (terminals/sessions) can drain at the same time without double-processing. Each item is a file claimed by an atomic rename, so concurrent claims are safe with no locks. Supports dynamic enqueue (`add`), a `--stop-file`/budget pause, and `reclaim` of a dead worker's stale claims. USE when work should be pulled by multiple concurrent workers from one queue ("process this queue from 3 terminals", "a shared work queue", "let several agents drain these in parallel safely"), or when items arrive over time. For a single orchestrator processing a fixed list, use /agentflow:foreach (its `--shard k/N` splits a list across terminals by static partition — queue is the dynamic, pull model instead).
allowed-tools	Bash, Read, Write, Edit, Agent
argument-hint	(--items <json> \| --checkbox <md> \| --folder <dir> \| --source <spec>) --prompt "<operation>" [--stop-file <path>] [--max-usd N] [--max-retries N]

/agentflow:queue

Make it visible: say in one line that you're draining a queue (id + how many pending).

A queue lives at .agentflow/queue/<id>/ as files moving through pending/ → claimed/ → done|failed/. A claim is an atomic rename, so any number of workers can pull from the same queue concurrently and never get the same item twice. You are one worker; the user may launch several (each its own terminal/session running this skill on the same <id>).

1. Create the queue (once)

node "${CLAUDE_PLUGIN_ROOT}/dist/state/queue.js" init <id> \
  (--items <file> | --checkbox <md> | --folder <dir> | --source '<json>') --prompt "<operation>" \
  [--stop-file <path>] [--max-usd N] [--max-retries N] [--force]

If the queue already exists (another worker created it), skip init and go straight to the drain loop.

2. Drain loop (each worker repeats until empty)

Claim the next item atomically:
```
node "${CLAUDE_PLUGIN_ROOT}/dist/state/queue.js" claim <id> [--worker <name>]
```
- {item: null} → the queue is empty (or paused: true via stop-file/budget) → stop, report.
- {item: {...}, task_prompt} → you claimed it; no other worker can.
Process the item (dispatch one Agent, or do it inline) using task_prompt + the item's data.

Commit:

node "${CLAUDE_PLUGIN_ROOT}/dist/state/queue.js" complete <id> <item-id> [--result '<json>']
# or, on failure:
node "${CLAUDE_PLUGIN_ROOT}/dist/state/queue.js" fail <id> <item-id> --error "…" [--retry]

Record agent token usage with budget-add so a --max-usd cap is real.

Loop to step 1. When claim returns {item:null}, the queue is drained (the run is marked done once pending and claimed both reach 0).

The Stop hook auto-resumes this loop across turns (per worker) until the queue is empty.

Operating notes

Many terminals: launch /agentflow:queue … (same <id>) in each — the rename-claim makes it safe. /agentflow:board shows the queue; queue status <id> shows pending/claimed/done/failed.
Dynamic work: queue add <id> --items <file> (or --source) enqueues more at any time; a drained queue reopens.
Pause: touch <stop-file> → every worker's next claim returns nothing and the Stop hook won't resume; remove it to continue. A budget cap (--max-usd) pauses the same way.
Dead worker: queue reclaim <id> --older-than <seconds> returns items stuck in claimed/ (older than N seconds) back to pending/ so another worker can pick them up.

queue

Informações da origem

Opções de instalação

Revise os arquivos de origem

Mais deste repositório

Mais deste repositório

/agentflow:queue

1. Create the queue (once)

2. Drain loop (each worker repeats until empty)

Operating notes

/agentflow:queue

1. Create the queue (once)

2. Drain loop (each worker repeats until empty)

Operating notes