Skip to content

Guide · Agent skill

How to use the Set Up Code Maintenance Bot agent skill

Updated · written with AI from the project's own documentation and checked against our catalogue data

The setup-code-maintenance-bot skill is an agent skill that guides an AI agent through creating a scheduled code maintenance bot inside a Paperclip instance, a platform for orchestrating agents. This guide explains what the skill does, how to install it, how to trigger it with a plain request, and what you need in place before you let it write to a repository. By the end you will know how to set up a bounded, paused routine that opens pull requests for your team to review.

Set Up Code Maintenance Botpaperclipai/paperclip · MIT

Quick start

  1. 1Install the skill. Download the ZIP from this page and unzip it into ~/.claude/skills/, or into .claude/skills/ inside a project. In the Claude apps, upload it under Settings → Capabilities → Skills.
  2. 2Prepare your Paperclip access. Confirm you have a Paperclip instance, an account that can create agents, skills and routines, and API access or the web UI.
  3. 3Gather repository and AI credentials. Have Git credentials that let the bot read, push and open pull requests, plus an AI connection for the agent's model provider.
  4. 4Ask for the setup in plain language. Describe the maintenance outcome, the repository and the schedule. Answer the agent's questions about instance, company and agent.
  5. 5Choose the maintenance skill and limits. Select the maintenance skill, timezone, time, run limit and maximum PRs per run and outstanding PRs.
  6. 6Review readiness and enable. Check the summary and readback. Enable the routine only after readiness checks pass and you approve scheduling.

What the setup-code-maintenance-bot skill does

An agent skill is a folder of instructions and reference files that an AI agent loads when a task matches the skill's description. This one covers setup only: it asks which Paperclip instance, company, agent, repository, maintenance skill, schedule and limits to use, then builds a routine that runs a chosen maintenance task on a repository.

The maintenance judgment itself comes from a separate maintenance skill that you select, such as test auditing, documentation upkeep, dependency updates or refactoring. Once the routine is active, the bot can periodically apply that task and open pull requests for review.

Install the skill from the ZIP download

Download the skill as a ZIP file from this page and unzip it. For a personal setup, place the unzipped skill folder in ~/.claude/skills/. For a single project, place it in that project's .claude/skills/ folder instead.

If you use the Claude apps rather than a local folder, open Settings, go to Capabilities, and upload the skill under Skills.

  • Personal: ~/.claude/skills/
  • Project only: .claude/skills/ inside the project
  • Claude apps: Settings → Capabilities → Skills
setup-code-maintenance-bot/ agents/openai.yaml references/paperclip-api.md references/run-instructions.md SKILL.md

Requirements before you start

The skill only configures Paperclip, so it cannot run on its own. Gather these items before you ask the agent to set anything up. The skill runs network calls and needs credentials for repository access.

  • A Paperclip instance and an account with permission to create agents, skills and routines
  • Access to a Git repository the bot can read, push to and open pull requests against, with the credentials needed for that
  • An AI connection or account for the agent's model provider
  • Paperclip API access or the Paperclip web UI

Trigger the skill with a natural request

You do not need a special command. Describe the outcome you want in plain language and name the repository and schedule. The agent should then ask for any choices it still needs, such as the instance, the agent and the timezone.

If you only want a plan, say so. The skill is written to produce a setup plan and stop before it provisions or runs anything when asked only to draft instructions.

  • Set up a nightly maintenance bot for our backend repo that keeps tests audited and opens PRs for review.
  • Draft a setup plan for a weekly docs upkeep bot. Don't create anything yet.
  • Create a scheduled dependency update routine for this repository, paused until I approve it.

What the setup walks through

The skill works in ordered stages. It verifies the Paperclip instance link and company, then asks you to choose an existing agent or create a new one. It resolves the repository, target branch, skill source, trigger, limits and PR delivery rules, and checks for existing routines so it does not create duplicate scheduled jobs.

The routine is created paused with its trigger disabled. The skill enables it only after readiness checks pass and scheduling is authorized. It then reads back the saved schedule and next run time.

  • Pick the Paperclip instance and company, then the agent
  • Set repository, branch, maintenance skill, timezone and time
  • Set limits such as run time, maximum PRs per run and outstanding-PR cap
  • Review the summary, then enable the routine only when ready

Safety notes: network access and write permissions

The skill uses the network and does not run scripts. It proposes no automatic merging, but it does grant write access to repositories and pull request credentials. Review every change the bot opens before you grant that access, and keep the routine paused until you are satisfied.

The skill is designed to keep the routine paused while any hire approval, credential, workspace or skill is unresolved. It also does not change shared project defaults to make the bot work.

  • No scripts are run by the skill.
  • Network access is used for Paperclip and repository operations.
  • Review pull requests as you would any contribution from a new team member.
  • Use a limited-scope credential for the bot where your Git host allows it.

Tips for a smoother setup

Name the timezone explicitly, because the skill will not infer one. Start with one coherent pull request per run and a small outstanding-PR cap, then raise limits once you trust the output.

If the Paperclip API is unavailable and browser access works, the skill can perform the same setup through the Paperclip web UI, inspecting existing records first to avoid duplicates.

  • Give a concrete time and timezone, for example a nightly run in your team's local zone.
  • Keep one routine per repository and task to avoid duplicate jobs.
  • Check the routine's next run time after enabling it.
  • Choose a maintenance skill whose scope matches the work you want.

What people use it for

  • A nightly bot that audits tests and opens one reviewable pull request per run.
  • A weekly documentation upkeep routine that proposes doc fixes for maintainers to review.
  • A scheduled dependency maintenance routine that opens update pull requests, capped at a set number of open PRs.
  • A refactoring routine that works on a defined area of a repository with stated exclusions.

Frequently asked questions

How do I install the setup-code-maintenance-bot skill?

Download the skill ZIP from this page and unzip it. Put the folder in ~/.claude/skills/ for personal use or in a project's .claude/skills/ folder for one project. In the Claude apps, you can instead upload it under Settings → Capabilities → Skills.

Is the setup-code-maintenance-bot skill free to use?

The skill is published under the MIT licence, as listed on its page. Using it still requires a Paperclip instance, repository access and an AI account for the agent's model provider, which may have their own costs.

What do I need to run the code maintenance bot?

You need a Paperclip instance and an account that can create agents, skills and routines. You also need Git access the bot can read, push to and open pull requests against, an AI connection for the agent, and either Paperclip API access or the web UI.

Does the bot merge pull requests automatically?

No. The skill proposes no automatic merging. It opens pull requests for your team to review, but it does grant write access to repositories and PR credentials, so review each change before granting access.

Can I use this skill without Paperclip?

No. The skill only configures a Paperclip instance and cannot run standalone. If you do not use Paperclip, this skill does not apply to your setup.

What if the skill's setup fails or the API is not working?

If the Paperclip API is unavailable and browser access is available, the skill can complete the same setup through the Paperclip web UI. It inspects existing records first so it does not duplicate resources, and it keeps the routine paused until readiness is confirmed.

New agent skills, repos and AI tools — every week

The best new skills and open-source projects, the fastest-growing repos, and one free prompt. Five minutes to read. Unsubscribe anytime.

Our verdict

This skill suits engineering teams that run Paperclip already and want scheduled maintenance pull requests with a bounded, inspectable routine and human review. It is not a fit if you want unattended merges, do not use Paperclip, or expect the setup skill itself to judge code quality, since that comes from the separate maintenance skill you choose.

Download the skill