flue-workflows

flue-workflows is a skill for Claude Code from pandysp/claude-plugins. It costs 141 tokens per session (2,933 once invoked), scanned A, original, MIT.

A tool for writing and running JavaScript programs that coordinate multiple Flue workers, which are separate workers handling parts of a task. It supports parallel work, independent checks, conditional follow-up, crash-resumable runs, and retained changes in separate Git copies.

In plain words
What is it for?
It is for building multi-agent pipelines, fan-out research, per-item processing, parallel coding tasks, independent reviews, and workflows that resume after a crash.
Why use it?
It helps when a task is too broad or long for one agent and needs independent workers, branching decisions, or work that can continue after interruption.

Skill for Claude Code

Written for Claude Code: shipped in a Claude Code plugin. Also seen: mentions subagents; mentions Claude Code; mentions Codex.

Needs its repository: it runs a file that does not travel with it, so clone the repository first. The line is node /absolute/workflow-space/flue.mjs doctor --model openai-codex/gpt-5.5 --auth pi.

Part of the flue-workflows plugin — 1 skill shipped together

Good fit It is for building multi-agent pipelines, fan-out research, per-item processing, parallel coding tasks, independent reviews, and workflows that resume after a crash.

Compare 6 skills from other repositories ↓
Install

Getting it into your agent

It runs from inside its repository, so the clone comes first — what it calls does not travel with the file alone.

Clone the repo
git clone --depth 1 https://github.com/pandysp/claude-plugins
agentmods
npx agentmods add skills/pandysp/claude-plugins/flue-workflows

Made for: Claude Code.

Or install flue-workflows, the plugin that ships this one along with the rest of its 1 skill.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for flue-workflows

README.md
[![agentmods](https://agentmods.dev/badge/skills/pandysp/claude-plugins/flue-workflows/github.svg)](https://agentmods.dev/skills/pandysp/claude-plugins/flue-workflows)
Your own site
<a href="https://agentmods.dev/skills/pandysp/claude-plugins/flue-workflows"><img src="https://agentmods.dev/badge/skills/pandysp/claude-plugins/flue-workflows/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

agentmods 80×15 button for flue-workflows

Your own site · 80×15
<a href="https://agentmods.dev/skills/pandysp/claude-plugins/flue-workflows"><img src="https://agentmods.dev/badge/skills/pandysp/claude-plugins/flue-workflows.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 141 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 2,933 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
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 5.1 $0.00141 $0.02933
Opus 5.5 $0.00056 $0.01173
Sonnet 5.5 $0.00028 $0.00587
Haiku 4.5 $0.00014 $0.00293

Measured 6d ago against content hash 53342e7de1d4, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade A, and why

flue-workflows 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 6d ago.

The scan reads SKILL.md. This mod also ships 27 executable files (examples/coding/create-fixture.mjs, examples/coding/program.mjs, examples/coding/support.mjs, …), listed below but not scanned — reading those needs a real analyzer, not pattern matching.

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.

plugins/flue-workflows/skills/flue-workflows/SKILL.md · 233 lines

How it starts

The opening of the file, as written. The whole thing — 233 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Flue workflows

Write a program when the next work depends on what earlier workers find. Keep ordinary judgment in the main conversation. The program supplies the control flow; Flue supplies the workers. Do not hand the overall task to a supervisor model or replace it with a predefined workflow catalog.

When to use it

Use it when the user asks for Flue workflows, or when the work needs something the host's own subagents or workflow tool cannot give: workers on another provider or subscription, a long fan-out that must survive a crash and resume (within its --timeout), or parallel coding workers in independent Git copies with retained patches. Otherwise prefer the host's mechanism: it needs no setup and no unrestricted local workers.

Start with the work, not a fleet

Scout enough to name the inputs, outputs and decisions. Decide which checks actually need independence and where results must meet. Use ordinary JavaScript for loops, branches, deduplication, voting and stopping rules. Worker roles and result shapes can be invented per call; none must be selected from a catalog.

Before running, establish:

  • Authority: workers run local commands and can access the whole host. Obtain explicit unrestricted-access approval. No containers or command approval UI are provided. A private Git copy is not security containment.
  • Inputs: name a working directory and put programs in a separate directory. Keep the workflow workspace outside those directories. Do not snapshot a broad home/tmp directory or put changing run files among program inputs.
  • Credentials/model: choose explicitly; the authoring host's model, login, instructions and tools are not inherited. --auth pi reads an existing valid OpenAI or Claude subscription login from pi; --auth env:VARIABLE deliberately selects API-key billing (credential sources). Never print, copy or refresh credentials to make a test work.
  • Coverage: define complete, partial and failed outcomes; report discovered, attempted, failed and deliberately omitted work. A finished program, valid JSON or a vote count does not prove the answer is true or complete.
  • What workers can see: workers cannot see images. Flue's read tool returns a PNG or JPEG as garbled text, with no error. Do not ask workers to check rendered pages, screenshots or charts; look at them in the main session. Drop this note once flue#863 is fixed.

Read the full file on GitHub · 233 lines

Files

What ships with it

33 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.

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. 6d ago Changed · +4 lines 53342e7de1d4
  2. 25d ago First seen · 229 lines · 141 tokens per session scan A eca0d846761c

Subscribe to this mod's changes

flue-workflows is a skill published in the GitHub repository pandysp/claude-plugins (6 stars, last pushed 7d ago), licensed MIT. It adds 141 tokens to every session and 2,933 once invoked, about $0.0006 per session on Opus 5.5. 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 2026-10-03.