Agent skill

Logical Commits

by TotalCross in TotalCross/totalcross

Create focused Git commits that follow this repository's validated commit-message format, with required scopes, focused validation, and preservation of unrelated local changes; invoke only when the…

LGPL-2.1Auto-check passedDevelopment

Install Logical Commits

skills CLI
$ npx skills add TotalCross/totalcross --skill logical-commits -a claude-code

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

GitHub CLI
$ gh skill install TotalCross/totalcross logical-commits --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/TotalCross/totalcross.git skills-src && mkdir -p .claude/skills && cp -r skills-src/.agents/skills/logical-commits .claude/skills/logical-commits && 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
logical-commits
GitHub stars
227
Token cost
~1.3k tokens
SKILL.md length
625 words
Files
2
Skills in repo
2
Repo updated
First seen
Licence
LGPL-2.1

At a glance

Create focused Git commits that follow this repository's validated commit-message format, with required scopes, focused validation, and preservation of unrelated local changes; invoke only when the…

  • Works in 11 steps: Read the active ExecPlan state first… → Inspect scoped changes → Split changes by behavior or contract,… → …
  • An active ExecPlan explicitly requires commits
  • Instructions only: no scripts, shell commands, URLs or credentials in SKILL.md
  • Tasks that involve Commit messages

What it does

Logical Commits is an agent skill from TotalCross/totalcross. Create focused Git commits that follow this repository's validated commit-message format, with required scopes, focused validation, and preservation of unrelated local changes; invoke only when the user asks to commit or an active ExecPlan explicitly requires commits.

Its SKILL.md is about 1.3k tokens, which your agent loads only when the skill is triggered. The skill folder holds 2 other files (for example `agents/openai.yaml`).

It sits in Development, covering Commit messages. It works with Java, Git and Linux. The repository describes itself as: TotalCross is a Software Development Kit that helps cross platform application development. Currently supported platforms are: Windows, Wince, Android, iOS, Linux and Linux ARM…

When your agent uses it

  • An active ExecPlan explicitly requires commits
  • Tasks that involve Commit messages

Example prompts

  • “/logical-commits”

Requirements

  • Python 3

Workflow steps

11 steps, taken from the first numbered list in SKILL.md.

  1. Read the active ExecPlan state first when executing an ExecPlan. Determine
  2. Inspect scoped changes
  3. Split changes by behavior or contract, not by file count. Suitable commit boundaries include
  4. Before each commit, run the validate-headers skill and the smallest
  5. Stage only intended paths
  6. Write an English commit message in the repository's
  7. Before creating the commit, write the exact proposed message to a temporary
  8. Create the commit from the same validated message file
  9. Validate the created commit with the same repository validator
  10. Update the active ExecPlan state after the logical commit. Update the active
  11. Report the commit hash, subject, paths, focused validations, and any deferred

What it can do on your machine

Read from SKILL.md and the folder at commit 34cc7ee. 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

    No scripts in the folder and no shell commands in SKILL.md.

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

  • Network

    No URLs in SKILL.md.

    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

Logical Commits loads about 1.3k tokens when it runs. Until then it costs about 71 tokens; SKILL.md has 625 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~71
When it runs · the whole SKILL.md, loaded when a task matches
~1.3k

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 TotalCross/totalcross at commit 34cc7ee, republished under its LGPL-2.1 licence (© TotalCross). 625 words, ~1,297 tokens.

Download SKILL.mdSave it as .claude/skills/logical-commits/SKILL.md (or your agent's skills folder). This skill also uses 1 other file; get the full folder from GitHub.
name
logical-commits
description
Create focused Git commits that follow this repository's validated commit-message format, with required scopes, focused validation, and preservation of unrelated local changes; invoke only when the user asks to commit or an active ExecPlan explicitly requires commits.

Commit logical repository changes

This skill changes Git state. Confirm that committing is explicitly requested by the user or required by an active ExecPlan whose execution was requested.

  1. Read the active ExecPlan state first when executing an ExecPlan. Determine the current functional slice and intended paths. Follow .agent/PLANS.md when a plan is present.

  2. Inspect scoped changes:

    git status --short -- <task paths>
    git diff --stat -- <task paths>
    git diff -- <task paths>

    Inspect staged changes separately. Do not run destructive cleanup and do not include unrelated modifications.

  3. Split changes by behavior or contract, not by file count. Suitable commit boundaries include:

    • one shared interface plus its focused tests;
    • one platform-family migration;
    • one dependency scaffold and its documentation;
    • one release-policy change and its idempotence tests;
    • one follow-up fix discovered by validation.

    Do not create a commit for formatting-only fragments that belong to the same functional change. Do not combine architecture, unrelated cleanup, generated artifacts, and documentation for another feature.

  4. Before each commit, run the validate-headers skill and the smallest validation sufficient for that slice. Always run:

    git diff --check --cached

    Save verbose build output to a log and record only a compact summary.

  5. Stage only intended paths:

    git add -- <paths>

    Review:

    git diff --cached --stat
    git diff --cached -- <paths>
    
  6. Write an English commit message in the repository's Conventional-Commits-inspired format:

    <type>(<scope>): imperative description
    <type>(<scope>,<platform>): imperative description
    <type>!(<scope>[,<platform>]): imperative description
    <type>(<scope>[,<platform>])!: imperative description

    Use only these types: fix, feat, refactor, perf, style, test, docs, build, ci, chore, or revert. The scope is required and may contain letters, numbers, underscores, and hyphens. The optional platform must be lowercase and is the second qualifier. Architecture is not a title qualifier; mention it in the body instead.

    The title must be English, start with a lowercase letter, use the imperative mood, contain at least three words, be at most 80 characters, and not end with a period. If there is a body, separate it from the title with one blank line. Keep every body line at 80 characters or less; the CI check ignores Markdown link URLs when measuring body lines.

    For every non-trivial change, add a body that explains:

    • why the change is needed;
    • what behavior or contract changes;
    • platform, compatibility, artifact, or release impact;
    • focused validation completed and important deferrals.

    Example:

    refactor(build): centralize native target policy
    
    Resolve Android, Linux, and Windows platform settings from the shared
    native-build configuration so workflows and explicit target wrappers no
    longer repeat toolchain values.
    
    Keep minizip on Android API 24 while the default remains API 23. Validate
    configuration resolution and the zlib build-only workflow.

    For breaking changes, put ! in the title and add a BREAKING CHANGE: footer describing the compatibility impact. For revert commits, name the reverted commit hash or hashes in the body.

  7. Before creating the commit, write the exact proposed message to a temporary file and validate that file with the repository's source-of-truth validator:

    python3 .github/scripts/validate-commit-message.py \
      --message-file <message-file>

    Resolve every reported body-line warning before committing, even though CI currently reports those lines as warnings rather than failures. This avoids creating a commit that immediately requires a follow-up formatting repair.

  8. Create the commit from the same validated message file:

    git commit -F <message-file>

    Do not amend or rewrite history unless explicitly requested.

  9. Validate the created commit with the same repository validator:

    python3 .github/scripts/validate-commit-message.py --commit HEAD

    Do not maintain a second inline implementation of commit-message rules in this skill. The standalone validator is the source of truth shared with CI.

  10. Update the active ExecPlan state after the logical commit. Update the active plan only when this commit reaches a functional-family, architecture, ABI, release-policy, or milestone checkpoint.

  11. Report the commit hash, subject, paths, focused validations, and any deferred expensive validation. Do not push unless explicitly requested.

© TotalCross, LGPL-2.1. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

SKILL.md and 1 other file in .agents/skills/logical-commits of TotalCross/totalcross.

  • SKILL.md
  • agents/openai.yaml

Open the folder on GitHubat commit 34cc7ee

Compare with similar skills

Logical Commits 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.

Logical Commits compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Logical Commits this skillTotalCross/totalcross227—~1.3kAutomated safety check: PassLGPL-2.1
Commitmanhpham90vn/Deskhub133—~978Automated safety check: PassMIT
Neru Create PRy3owk1n/neru785—~1.9kAutomated safety check: PassMIT
Changelog Generator for Javadecebals/claude-code-java751—~2.1kAutomated safety check: PassMIT
Java Conventional Commit Writerdecebals/claude-code-java751—~1.4kAutomated safety check: PassMIT
Gpg SigningProrise-cool/Claude-Code-Multi-Agent305—~3.9kAutomated safety check: WarnNone

Similar skills

  • Commit

    manhpham90vn/Deskhub

    Write a Deskhub commit message. An agent skill from manhpham90vn/Deskhub.

    133 GitHub stars~978 tokensUpdated yesterday
    DevelopmentAuto-check passed
  • Neru Create PR

    y3owk1n/neru

    Commit working changes and open a Neru pull request the maintainer's way: conventional commit subjects, a PR title written for the changelog, the just ci gate, and the repo PR template filled…

    785 GitHub stars~1.9k tokensUpdated yesterday
    DevelopmentAuto-check passed
  • Changelog Generator for Java

    decebals/claude-code-java

    Builds changelog entries from conventional commits in a Java project, after working out whether it uses SemVer, two-part versions or calendar versions.

    751 GitHub stars~2.1k tokensUpdated 1 mo ago
    DevelopmentAuto-check passed
  • Java Conventional Commit Writer

    decebals/claude-code-java

    Writes Conventional Commits messages for Java projects from staged changes, with Java-specific types, scopes and a token-light workflow.

    751 GitHub stars~1.4k tokensUpdated 1 mo ago
    DevelopmentAuto-check passed
  • Gpg Signing

    Prorise-cool/Claude-Code-Multi-Agent

    Comprehensive guide to GPG commit signing. An agent skill from Prorise-cool/Claude-Code-Multi-Agent.

    305 GitHub stars~3.9k tokensUpdated 22 days ago
    Backend & APIsAuto-check: warnings
  • React Router Release Notes Prep

    remix-run/react-router

    Polishes pending React Router change files before the versioning scripts run, and decides whether a long-form What's Changed section is warranted.

    57k GitHub stars~1.1k tokensUpdated yesterday
    DevelopmentAuto-check passed

More from TotalCross/totalcross

  • Validate Headers

    TotalCross/totalcross

    Validate or fix LGPL-2.1-only headers, apply approved copyright provenance, and prepare provenance audits for renamed, split, copied, or extracted first-party code.

    227 GitHub stars~1.1k tokensUpdated today
    Auto-check passed

Works with

Categories

Questions about Logical Commits

What does Logical Commits do?

Create focused Git commits that follow this repository's validated commit-message format, with required scopes, focused validation, and preservation of unrelated local changes; invoke only when the…. Logical Commits is an agent skill from TotalCross/totalcross. Create focused Git commits that follow this repository's validated commit-message format, with required scopes, focused validation, and preservation of unrelated local changes; invoke only when the user asks to commit or an active ExecPlan explicitly requires commits.

When should I use Logical Commits?

Logical Commits fits situations like: an active ExecPlan explicitly requires commits; tasks that involve Commit messages.

How do I install Logical Commits in Claude Code?

Run `npx skills add TotalCross/totalcross --skill logical-commits -a claude-code`. Or copy the skill folder (.agents/skills/logical-commits in TotalCross/totalcross) into .claude/skills/logical-commits in your project. Claude Code loads it when a task matches its description.

How do I install Logical Commits in Codex?

Run `npx skills add TotalCross/totalcross --skill logical-commits -a codex`. Or copy the skill folder (.agents/skills/logical-commits in TotalCross/totalcross) into .agents/skills/logical-commits in your project. Codex loads it when a task matches its description.

Can I use Logical Commits 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 TotalCross/totalcross --skill logical-commits -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/logical-commits, .gemini/skills/logical-commits, .github/skills/logical-commits and .opencode/skills/logical-commits in your project.

What does Logical Commits need to run?

SKILL.md names no scripts, command-line tools or credentials: Logical Commits is instructions for the agent only. Our summary lists: Python 3.

Does Logical Commits access the network?

SKILL.md contains no URLs. Any network use would come from the scripts or tools the agent runs. This is read from the text; nothing was executed.

Is Logical Commits 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 Logical Commits use?

Logical Commits is published under the LGPL-2.1 licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Logical Commits use?

About 1.3k tokens (SKILL.md is roughly 5.2k 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 Logical Commits?

Skills that share tags, products or a category with Logical Commits: Commit (manhpham90vn/Deskhub, 133 stars), Neru Create PR (y3owk1n/neru, 785 stars), Changelog Generator for Java (decebals/claude-code-java, 751 stars) and Java Conventional Commit Writer (decebals/claude-code-java, 751 stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Logical Commits?

TotalCross (a GitHub organization) maintains it in TotalCross/totalcross, which has 227 GitHub stars. The repository holds 2 skills in this directory. The repository was last updated on October 8, 2026.

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