Agent skill

Finalize

by zifeo in zifeo/lade

Finalize a merge request: rebase onto main, run bugs, simplify, and specs in parallel, apply, then refactor last.

MPL-2.0Auto-check passedDevOps & Cloud

Install Finalize

skills CLI
$ npx skills add zifeo/lade --skill finalize -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI
$ gh skill install zifeo/lade finalize --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy
$ git clone --depth 1 https://github.com/zifeo/lade.git skills-src && mkdir -p .claude/skills && cp -r skills-src/.cursor/skills/finalize .claude/skills/finalize && rm -rf skills-src

Use ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.

Claude Code skills documentation · loads skills from .claude/skills/

Facts

Skill name
finalize
GitHub stars
133
Token cost
~1k tokens
SKILL.md length
574 words
Files
1
Skills in repo
2
Repo updated
First seen
Licence
MPL-2.0

At a glance

Finalize a merge request: rebase onto main, run bugs, simplify, and specs in parallel, apply, then refactor last.

  • Works in 5 steps: Sit on main first → Parallel: bugs, simplify, specs → Last: refactor → …
  • The user asks to finalize
  • SKILL.md covers When, 0. Sit on main first, 1. Parallel: bugs, simplify,… and 2. Last: refactor, plus 2 more sections
  • Calls git

What it does

Finalize is an agent skill from zifeo/lade. Finalize a merge request: rebase onto main, run bugs, simplify, and specs in parallel, apply, then refactor last. Use when the user asks to finalize, close a PR, or run the review pageant.

Its SKILL.md is about 1k tokens, which your agent loads only when the skill is triggered. It is a single SKILL.md file with no bundled scripts.

It sits in DevOps & Cloud, covering Refactoring and Secrets management. It works with Bash, Kubernetes and Rust. The repository describes itself as: Temporary access to secrets and private networks for one command, then gone. Same wrap for humans and agents. See which access was used. The licence is MPL-2.0.

When your agent uses it

  • The user asks to finalize
  • Run the review pageant

Example prompts

  • “/finalize”

Workflow steps

5 steps, taken from the step headings in SKILL.md.

  1. Sit on main first
  2. Parallel: bugs, simplify, specs
  3. Last: refactor
  4. Do not
  5. Report

What it can do on your machine

Read from SKILL.md and the folder at commit d775ef5. It shows what the files ask for, not the result of running them.

  • Tool permissions

    Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

    From allowed-tools in the SKILL.md frontmatter.

  • Runs code

    Shell commands in SKILL.md call:

    • git

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    No URLs in SKILL.md. Its commands use git, which can reach the network depending on how they are called.

    From URLs in SKILL.md, links to its own repository left out.

  • Credentials

    Names no API keys, tokens, secrets or passwords.

    From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Finalize loads about 1k tokens when it runs. Until then it costs about 49 tokens; SKILL.md has 574 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~49
When it runs · the whole SKILL.md, loaded when a task matches
~1k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check passed

The automated check found no risky patterns in SKILL.md.

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from zifeo/lade at commit d775ef5, republished under its MPL-2.0 licence (© zifeo). 574 words, ~1,019 tokens.

Download SKILL.mdSave it as .claude/skills/finalize/SKILL.md (or your agent's skills folder).
name
finalize
description
Finalize a merge request: rebase onto main, run bugs, simplify, and specs in parallel, apply, then refactor last. Use when the user asks to finalize, close a PR, or run the review pageant.
disable-model-invocation
true

Finalize

A pageant, not a swarm. Reviewers are read-only. The parent waits, edits, then launches the next seat. One pass.

When

User says /finalize, finalize, close the MR, or ready to merge. Do not run this on a drive-by question.

0. Sit on main first

Fetch origin/main. Rebase this branch onto it (or merge main if rebase is refused). If the rebase conflicts, stop and ask. Do not launch reviewers on a stale base.

Then pin what will be reviewed. Commits ahead of origin/main and uncommitted edits both count. A dirty tree is the review subject. Do not commit it to make a range. Stop only when both are empty.

bash
git fetch origin main
git merge-base origin/main HEAD
git diff --stat "$(git merge-base origin/main HEAD)"
git diff --stat HEAD

Build a Spec: block for the specs child, in this order:

  1. Specs the user stated in this conversation
  2. The PR body / issue, if this is an MR
  3. A file the user named as the spec

If none of those exist, still launch specs. It must ask. Do not drop the seat.

The child scope uses Diff: branch changes when the commit range is non-empty, and Diff: uncommitted changes when the work is only in the tree. When both exist, name both.

text
Full Repository Path: <abs>
Diff: <branch changes | uncommitted changes | both>
Base Branch: main
Range: <merge-base>...<HEAD>
Worktree: <clean | dirty, include HEAD>
Spec:
<quoted lines, or "none, ask">

Custom agent names may be missing on the Task tool. Default path: generalPurpose children. Paste .cursor/agents/<name>.md as the preamble, then the scope block.

1. Parallel: bugs, simplify, specs

Launch these three together. Wait for all three. Do not launch refactor here.

SubagentJobWrites
bugsConcrete failuresno
simplifySmaller form, same behaviorno
specsDiff vs a cited spec. Asks if no referenceno

If specs returns Ask:, answer from this conversation or ask the user. Then relaunch only specs once with the answer. Do not invent product.

Apply only if useful. Useful means a real failure, or a call site that should have used an existing helper for the same concern. Not useful: taste, a new generic, or a wc -l move that does not change how a human reads production code.

  1. bugs critical and high, if the failure is real.
  2. specs wrong / missing that cite a reference.
  3. simplify reuse of a named existing helper, or a dead wrapper with one caller. Do not extract a new helper as a leftover. Hunt the project, not only the hunk.

Narrowest check that proves the edit. Use the project's usual check. Do not widen permissions to make a build pass.

Show full SKILL.md (177 more words)Show less

2. Last: refactor

Only after step 1 is applied. Launch refactor alone.

SubagentJobWrites
refactorFiles over 300 (soft) / 350 (hard), including markdownno

Apply a hard file only if the split is two jobs a human can hold (production code). Moving tests to another file only to get under the line limit is not useful. Soft (301–350) is a proposal; apply only if the split is obvious and local. Keep the public API and test names.

Narrowest check again.

3. Do not

  • Commit, push, merge, or enable auto-merge
  • Request all for a compile
  • Launch refactor next to bugs
  • Let specs invent a contract
  • Mention source control unless the user asked

4. Report

One short status. Then leftovers as a table, only rows where Useful? is yes:

ItemUseful?Why

Useful? is yes only for a real failure or an existing helper to call. Skip taste, new helpers, and wc -l moves. A leftover is only a still-real failure or a named existing helper the diff still misses. "The reviewer mentioned it" is not a leftover.

© zifeo, MPL-2.0. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

Just SKILL.md in .cursor/skills/finalize of zifeo/lade.

Open the folder on GitHubat commit d775ef5

Compare with similar skills

Finalize next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

Finalize compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Finalize this skillzifeo/lade133—~1kAutomated safety check: PassMPL-2.0
Misemajiayu000/claude-skill-registry6661 repos~1.8kAutomated safety check: NotesApache-2.0
Niubashunixwin/niubash155—~2.1kAutomated safety check: PassMIT
Kopiur Designhome-operations/kopiur113—~2.4kAutomated safety check: PassAGPL-3.0
Flow Contextflowexec/flow137—~654Automated safety check: PassApache-2.0
Refactoratilladeniz/Kubeli387—~7.7kAutomated safety check: WarnMIT

Similar skills

  • Mise

    majiayu000/claude-skill-registry

    Manages per-project tool versions (Node.js, Python, Go, Rust, Ruby and hundreds more), environment variables and tasks from one mise.toml file, as a single binary that replaces nvm, pyenv, rbenv and…

    666 GitHub starsUsed in 1 repo~1.8k tokens
    DevOps & CloudAuto-check: notes
  • Niubash

    unixwin/niubash

    Run Windows tasks in Niubash, the GNU Bash-compatible Windows-native shell.

    155 GitHub stars~2.1k tokensUpdated 3 days ago
    DevelopmentAuto-check passed
  • Kopiur Design

    home-operations/kopiur

    Design norms and locked decisions for the Kopiur Kopia-native Kubernetes backup operator (Rust/kube-rs).

    113 GitHub stars~2.4k tokensUpdated today
    DevOps & CloudAuto-check passed
  • Flow Context

    flowexec/flow

    This project uses flow for automation. An agent skill from flowexec/flow.

    137 GitHub stars~654 tokensUpdated 6 days ago
    DevOps & CloudAuto-check passed
  • Refactor

    atilladeniz/Kubeli

    Refactors code following Ousterhout's design principles. An agent skill from atilladeniz/Kubeli.

    387 GitHub stars~7.7k tokensUpdated 2 days ago
    DevelopmentAuto-check: warnings
  • LangBot Deployment Guide

    langbot-app/LangBot

    Deploys and configures a LangBot instance with Docker Compose or Kubernetes, covering config.yaml, the Box sandbox runtime, the plugin runtime and the global API key.

    18k GitHub stars~1.2k tokensUpdated today
    DevOps & CloudAuto-check: notes

More from zifeo/lade

  • Lade Release

    zifeo/lade

    Prepare a Lade release: update CHANGELOG.md and bump crate versions with cargo set-version.

    133 GitHub stars~749 tokensUpdated today
    Auto-check passed

Questions about Finalize

What does Finalize do?

Finalize a merge request: rebase onto main, run bugs, simplify, and specs in parallel, apply, then refactor last. Finalize is an agent skill from zifeo/lade. Finalize a merge request: rebase onto main, run bugs, simplify, and specs in parallel, apply, then refactor last.

When should I use Finalize?

Finalize fits situations like: the user asks to finalize; run the review pageant.

How do I install Finalize in Claude Code?

Run `npx skills add zifeo/lade --skill finalize -a claude-code`. Or copy the skill folder (.cursor/skills/finalize in zifeo/lade) into .claude/skills/finalize in your project. Claude Code loads it when a task matches its description.

How do I install Finalize in Codex?

Run `npx skills add zifeo/lade --skill finalize -a codex`. Or copy the skill folder (.cursor/skills/finalize in zifeo/lade) into .agents/skills/finalize in your project. Codex loads it when a task matches its description.

Can I use Finalize in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add zifeo/lade --skill finalize -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/finalize, .gemini/skills/finalize, .github/skills/finalize and .opencode/skills/finalize in your project.

What does Finalize need to run?

Going by SKILL.md and its folder, Finalize needs the command-line tools its instructions call (git).

Does Finalize access the network?

SKILL.md contains no URLs. Its commands use git, which can reach the network depending on how they are called. This is read from the text; nothing was executed.

Is Finalize safe to install?

Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.

What licence does Finalize use?

Finalize is published under the MPL-2.0 licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Finalize use?

About 1k tokens (SKILL.md is roughly 4.1k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.

What are the alternatives to Finalize?

Skills that share tags, products or a category with Finalize: Mise (majiayu000/claude-skill-registry, 666 stars), Niubash (unixwin/niubash, 155 stars), Kopiur Design (home-operations/kopiur, 113 stars) and Flow Context (flowexec/flow, 137 stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Finalize?

zifeo (a GitHub user) maintains it in zifeo/lade, which has 133 GitHub stars. The repository holds 2 skills in this directory. The repository was last updated on October 7, 2026.

Source: zifeo/lade on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.