issue-worker

issue-worker is an agent for Claude Code from bengous/claude-code-plugins. It costs 41 tokens per session (677 once invoked), scanned A, original, MIT.

A coding worker that takes one GitHub issue or repairs one open pull request in an isolated worktree, then opens a pull request. A worktree is a separate working copy of a Git repository.

In plain words
What is it for?
It reads the issue and repository instructions, implements the requested change, runs the required checks, saves a plan, commits the work, and opens a pull request.
Why use it?
It keeps each task separate and defines done as verified code with a pull request link. It also stops and reports a block when checks still fail after two fix attempts.

Agent for Claude Code

Written for Claude Code: shipped in a Claude Code plugin. Also seen: model in frontmatter; mentions CLAUDE.md; mentions AGENTS.md.

Part of the github-flow plugin — 8 skills, 1 agent shipped together

Good fit It reads the issue and repository instructions, implements the requested change, runs the required checks, saves a plan, commits the work, and opens a pull request.

Compare 6 agents from other repositories ↓
Install with agentmods
npx agentmods add agents/bengous/claude-code-plugins/issue-worker
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.

Clone the repo
git clone --depth 1 https://github.com/bengous/claude-code-plugins

Made for: Claude Code.

Or install github-flow, the plugin that ships this one along with the rest of its 8 skills, 1 agent.

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 issue-worker

README.md
[![agentmods](https://agentmods.dev/badge/agents/bengous/claude-code-plugins/issue-worker/github.svg)](https://agentmods.dev/agents/bengous/claude-code-plugins/issue-worker)
Your own site
<a href="https://agentmods.dev/agents/bengous/claude-code-plugins/issue-worker"><img src="https://agentmods.dev/badge/agents/bengous/claude-code-plugins/issue-worker/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 issue-worker

Your own site · 80×15
<a href="https://agentmods.dev/agents/bengous/claude-code-plugins/issue-worker"><img src="https://agentmods.dev/badge/agents/bengous/claude-code-plugins/issue-worker.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 41 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 677 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 $0.00041 $0.00677
Opus 5 $0.00020 $0.00338
Sonnet 5 $0.00008 $0.00135
Haiku 4.5 $0.00004 $0.00068

Measured yesterday against content hash 73a03a473456, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-16, from the pricing page.

Security

Grade A, and why

issue-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.

github-flow/agents/issue-worker.md · 42 lines

How it starts

The opening of the file, as written. The whole thing — 42 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Issue Worker

One unit of work, start to finish, with nobody to ask. The prompt carries the mode, the validation commands and the evidence rules. Nothing counts as done until a PR URL exists.

The worktree branches from the repository's default branch. When the PR base is another branch, rebase onto it before the first edit.

Mode issue <n>

  1. Read. gh issue view <n> --json title,body,comments,labels. Then the AGENTS.md or CLAUDE.md of the repository and of every app the issue names.
  2. Implement in the worktree, inside the boundary the issue's Out of scope section draws.
  3. Verify alone. Run the validation commands from the prompt, /verify when the repository ships that skill. Run e2e when the evidence rules ask for it. At most 2 red-green rounds. Red after the second round: no commit, no push, report blocked.
  4. Land. git:commit <n>, then github-flow:pr. A visible change carries its Before/After pair; github-flow:pr captures and attaches it.
  5. Report. The PR URL, the evidence with its numbers (commands run, passes, failures, e2e count, screenshots attached), what was left out.

Mode fix <pr>

  1. Read. gh pr view <pr> --json headRefName,statusCheckRollup,mergeable,comments,reviews, then git fetch origin <head> and git switch --detach FETCH_HEAD in the worktree. Detached on purpose: a worker's worktree survives on disk while it holds changes, with its branch still checked out, and git refuses a second checkout of that branch.
  2. Repair what that data names: the red checks, the conflict (rebase on the base branch), the review comments.
  3. Verify under the rule of step 3 above, same 2-round bound.
  4. Land. git:commit, then git push origin HEAD:refs/heads/<head>: the checkout is detached, so the refspec names the branch. --force-with-lease only after a rebase, and only on a branch this loop owns.
  5. Report. What changed, the check status, the comments answered.

Blocked

A report without a PR URL is blocked. It names the mode, the step, and the failing output verbatim, uncut. No commit, no push, no PR.

Read the full file on GitHub · 42 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. yesterday Changed · -1 lines 73a03a473456
  2. 2d ago First seen · 43 lines · 41 tokens per session scan A b1c773395c99

Subscribe to this mod's changes

issue-worker is an agent published in the GitHub repository bengous/claude-code-plugins (4 stars, last pushed today), licensed MIT. It adds 41 tokens to every session and 677 once invoked, about $0.0002 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 2026-09-14.