agentleFS
quatico-solutionsquatico-solutions/agent-skills/skills/commit-notation/SKILL.md

commit-notation

Skill2 starsChanged 4 months agoMITView on GitHub

Format git commit messages using Quatico's notation (Arlo's v1 with extensions). Use when: committing changes, making commits, creating commits, writing commit messages, reviewing commits, or analyzing git log. Triggers: commit, make commit, create commit, git commit, commit this, commit message, commit notation, Arlo notation.

The file

We split this skill into the rules it gives an agent

4 rules. About 0 are common, 3 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈3

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Validated (UPPERCASE): Test-verified - ran tests, all pass.

    line 49
  2. Very specific:

    Check branch name: Look for pattern like XXX-123

    line 113
  3. Distinctive:

    Check recent commits: If not in branch name, check commits not yet merged to main:

    line 118
  4. Distinctive:

    Ask user: If still not found, ask:

    line 123

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Commit Notation

Source of truth: QuaticoCommitNotation (GitHub) SYNC REQUIRED: Changes here MUST stay consistent with the canonical repo. Changes incomplete without sync.

Quatico uses Arlo's Commit Notation v1 with extensions (T, E).

Quick Reference

Risk LevelFormatExample
Known Safexr Extract method getData
ValidatedXR: Extract method getData
RiskyX!!R!! Rename across files
BrokenX**F** WIP: new feature

Intentions

Core

LetterNameUse When
FFeatureAdding or changing behavior
BBugfixFixing broken behavior
RRefactoringChanging implementation, not behavior
DDocumentationNon-code documentation changes

Extensions

LetterNameUse When
TTest-onlyAlter tests without altering functionality
EEnvironmentDev setup, tooling, non-code changes
WIPUnknownMixed changes, just checking in

Risk Level Guidelines

Known Safe (lowercase): Provable correctness - tool-verified refactoring, type-checked renames. Validated (UPPERCASE): Test-verified - ran tests, all pass. Risky (XX!!): Attempted but not fully verified. Broken (XX)**: WIP, may not compile, no verification.

Commit Message Format

X: Summary in active voice

TICKET-123

Rules:

  • Separator: : or nothing (e.g., F: Add feature or F Add feature)
  • Voice: Active ("Add feature") not progressive ("Adding feature")
  • Length: Subject < 75 chars
  • Ticket: Last line, blank line before, NO # prefix (git treats # as comment)

Writing Commit Messages

  1. Determine intention: F, B, R, D, T, E, or WIP
  2. Assess risk level: How verified is this change?
  3. Find ticket number: See below
  4. Format: [prefix] Summary with ticket on last line

Keep changes small: Features/bugfixes should be ≤8 LoC for Validated level.

What the subject should name

A fix names the defect. A feature names the outcome.

These pull in opposite directions, and applying one rule to both produces euphemism. The outcome of a fix is "it works again", which tells a reader nothing.

WeakStrong
BB: Unblock the build that the new check stoppedB: Stop the license check failing every build on the docs package
BB: Ship types a consumer can compile againstB: Export the parser types that consumers need to type-check
FF: Add a script that writes the exports mapF: Generate the exports map from the barrel directories

The tell in a weak fix subject is an abstract subject carrying a relative clause — "unblock the build that …", "ship the files that …". It reports that something improved without naming what was broken.

Someone scanning git log for the cause of a red build should find it in a subject line. Name what fails, where, and how widely.

Quick Checklist

Copy and track progress:

  • Intention identified (F/B/R/D/T/E/WIP)
  • Single intention per commit
  • Risk level assessed (lowercase/UPPER/!!/**)
  • Active voice summary
  • A fix names the defect; a feature names the outcome
  • Subject < 75 chars
  • Ticket on last line (if applicable)

Ticket Numbers

Finding the Ticket

  1. Check branch name: Look for pattern like XXX-123
    • FOO-123-my-bugfix → FOO-123
    • bugfix/FOO-123-description → FOO-123
    • feature/BAR-456-new-feature → BAR-456
  2. Check recent commits: If not in branch name, check commits not yet merged to main:
    git log origin/main..HEAD --oneline
  3. Ask user: If still not found, ask:
    • Is this an infrastructure change (leave empty)?
    • What's the ticket number?

When to Omit Ticket

Infrastructure changes may omit tickets:

  • CI/CD pipeline fixes
  • Development environment setup
  • Release version bumps
  • Dependency updates
  • Code formatting/linting

Examples

F: Add user authentication endpoint

FOO-123
B!! Fix race condition in event handler

BAR-456
r Extract method calculateTotal

FOO-123
D: Update CI pipeline documentation

See REFERENCE.md for full notation details and risk level criteria.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.