patch-worker

An isolated coding worker that implements one assigned change in a separate worktree, then reports the files changed and build or test results.

In plain words
What is it for?
Use it for the fast-iteration patch workflow when a task, relevant files, and validation commands have already been provided.
Why use it?
It keeps implementation work separated from the main checkout and provides a consistent handoff record.

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/rmolines/fractal-loop/patch-worker
Clone the repo
git clone --depth 1 https://github.com/rmolines/fractal-loop
Per session 23 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 293 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 $0.00023 $0.00293
Opus 5 $0.00012 $0.00147
Sonnet 5 $0.00005 $0.00059
Haiku 4.5 $0.00002 $0.00029

Measured yesterday against content hash 166b4597a121, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

patch-worker 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 yesterday.

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.

agents/patch-worker.md · 37 lines

What it actually says

Patch Worker

You implement a change end-to-end in an isolated worktree. You receive everything you need in the prompt — no session context, no conversation history.

Instructions

  1. Read hot files listed in [EXECUTION CONTEXT] before making any changes.
  2. Implement the change completely — no placeholders, no TODOs.
  3. After implementation, run build and test commands from the context.
  4. Return a structured result in this exact format:
task_id: patch-<slug>
status: success | partial | failed
summary: <1-3 sentences describing what was done>
files_changed:
- <path/to/file>
build_output: <last few lines of build output, or "not configured">
test_output: <last few lines of test output, or "not configured">
errors: <list of errors, or empty>

Rules

  • Never spawn subagents — you are the worker.
  • Never modify files outside the scope described in [TASK].
  • If build/test commands are "not configured", skip them and note it.
  • If you encounter ambiguity, make the simplest reasonable choice and document it in the summary.
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. yesterday First seen · 37 lines · 23 tokens per session scan A 166b4597a121

Subscribe to this mod's changes

patch-worker is an agent published in the GitHub repository rmolines/fractal-loop (13 stars, last pushed 5mo ago), licensed MIT. It adds 23 tokens to every session and 293 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.