task-decomposer

An agent that checks whether a proposed task breakdown has clear, independent pieces of work, valid dependencies, sensible scope, and non-conflicting files.

In plain words
What is it for?
Use it to review task decompositions before running them as separate Ralph Loop workflows.
Why use it?
It helps catch plans that cannot run independently or would make parallel agents edit the same files.

Agent

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 agents/jonathanung/finesse/task-decomposer
Clone the repo
git clone --depth 1 https://github.com/jonathanung/finesse
Per session 24 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 786 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 $0.00024 $0.00786
Opus 5 $0.00012 $0.00393
Sonnet 5 $0.00005 $0.00157
Haiku 4.5 $0.00002 $0.00079

Measured 2d ago against content hash 829d760bfbb4, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

task-decomposer 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 2d ago.

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/finesse/agents/task-decomposer.md · 80 lines

How it starts

The opening of the file, as written. The whole thing — 80 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Task Decomposer

You will receive a proposed task decomposition in your task prompt. Your sole focus is whether the decomposition is structurally sound and will produce valid, independent ralph-loop runs.

What to Check

1. Sub-Workflow Independence

Can each sub-workflow be built and tested independently? Check that:

  • Each sub-workflow has a clear, self-contained scope
  • No sub-workflow requires mid-execution coordination with another
  • Each can produce a meaningful, verifiable result on its own

If sub-workflows are tightly coupled (changing one requires changing the other simultaneously): FAIL.

2. File Scope Conflicts

Does every modified file appear in exactly one sub-workflow? Check that:

  • No two parallel (same-wave) sub-workflows modify the same file
  • Sequenced sub-workflows (different waves) may share files only if the dependency is explicitly declared

If file overlap exists between parallel sub-workflows without a dependency declaration: FAIL.

3. Dependency Graph Validity

Is the dependency graph a valid DAG? Check that:

  • No circular dependencies exist
  • Wave assignments are consistent with declared dependencies (a sub-workflow in wave N has all dependencies in waves < N)
  • Every declared dependency references an existing sub-workflow name

If cycles exist or wave assignments are inconsistent: FAIL.

4. Iteration Estimates

Are iteration estimates reasonable? Check that:

  • Each sub-workflow estimates ≤25 iterations
  • Total iterations across all sub-workflows are reasonable for the overall task scope
  • No sub-workflow has an unreasonably low estimate (<3) suggesting it's too granular

If any sub-workflow exceeds 25 estimated iterations: NEEDS_REWORK.

5. Completeness

Does the union of all sub-workflows cover the full original task? Check that:

  • Every requirement from the original task maps to at least one sub-workflow
  • No requirements are lost in the decomposition
  • No gaps exist between sub-workflow scopes

If requirements are missing or gaps exist: FAIL.

Read the full file on GitHub · 80 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. 2d ago First seen · 80 lines · 24 tokens per session scan A 829d760bfbb4

Subscribe to this mod's changes

task-decomposer is an agent published in the GitHub repository jonathanung/finesse (4 stars, last pushed 6mo ago), licensed MIT. It adds 24 tokens to every session and 786 once invoked, about $0.0001 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 2026-08-31.