agentleFS
pltnkpltnk/conventional-comments-skill/SKILL.md

conventional-comments

Skill1 starsChanged 6 months agoMITView on GitHub

Format review and feedback comments using the [Conventional Comments standard](https://conventionalcomments.org/). The goal is comments that are easy to understand and act on, with a collaborative tone. Use whenever the user asks for a review, critique, or feedback on anything — code, docs, RFCs, designs, drafts, PRs, or any other artifact — including indirect prompts like "thoughts on this" or "what could be better?". Defines comment format and tone, not review content, structure, or length. Labels and decorations in English; subject and discussion in the user's language.

3 other projects use a version of this.

Adopted by

Most recently taken up by

  • Other versions3 projects use a version of this.

    Word for word · 0

    None found.

    Edited · 0

    None found.

    Wrote their own version · 3

    1. dgalarza/agent-skills★ 72same name
    2. milistu/agent-skills★ 8same name
    3. roziscoding/conventional-comments★ 0same name

    How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.

The file

We split this skill into the rules it gives an agent

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

  • 0

    Common

    Most of the 3 edited versions we read keep these.

  • 3

    Distinctive

    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:

    Every comment follows this shape:

    line 16
  2. Distinctive:

    Use one label per comment.

    line 31
  3. Very specific:

    Don't assume full context.

    line 64
  4. Distinctive:

    Don't fragment one concern across many tiny comments.

    line 86

“Common” is measured: kept by at least 6 in 10 of the 3 edited versions of this file we could read. “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

Scope

This skill defines the format of each review comment and recommended communication practices. It does not define:

  • Review content. Substance, depth, and what's worth flagging come from the artifact and the user's request, not from this skill.
  • Review structure. This skill formats individual comments; it doesn't organize them — pick the structure you find the most suitable.
  • Review length. The number of comments must be as much or as little as the work warrants.

Format

Every comment follows this shape:

<label> [decorations]: <subject>

[discussion]
  • label — a single lowercase English word signaling the kind of comment. Bold in Markdown (e.g. **suggestion:**).
  • decorations (optional) — lowercase comma-separated qualifiers in parentheses, sitting between label and colon. Decorations classify comments that share a label. Example: **issue (ux,non-blocking):**.
  • subject — the main message, on the same line as the label.
  • discussion (optional) — supporting context, reasoning, examples, patches, or "why / next steps" notes. When a discussion is present, a blank line between the subject line and the discussion is required.

Labels

  • Use one label per comment.
  • Diverging from the following list is fine if a different label fits better.
LabelMeaning
praise:Highlights something genuinely positive. Aim for at least one per review; never invent it.
nitpick:A trivial preference-based request. Non-blocking by nature.
suggestion:Proposes an improvement. Be explicit about what and why. Consider patches and a (blocking) or (non-blocking) decoration.
polish:Like suggestion:, but for purely quality-improving changes — nothing is wrong.
issue:A specific problem you're confident about. Always pair with a separate follow-up suggestion: that proposes the fix; never embed the fix in the issue: discussion. If you're not sure a problem is real, use question: instead of issue:.
todo:A small, trivial, but necessary change. Lighter weight than issue: or suggestion:.
typo:Like todo:, but specifically for misspellings.
question:A potential concern where it isn't yet clear whether it's relevant. Ask the author for clarification.
thought:An idea that surfaced during review. Non-blocking by nature.
chore:A simple task required before acceptance (running a CI job, signing off a checklist). Link to the process when possible.
note:Always non-blocking. Highlights something to be aware of.

Decorations

DecorationMeaning
(non-blocking)Does not block acceptance.
(blocking)Blocks acceptance until resolved.
(if-minor)Resolve only if doing so turns out to be minor or trivial.

Decorations can be review-specific — (security), (ux), (test), (perf), (a11y) are common. Use them when they sharpen the comment. Keep the vocabulary of decorations small and unambiguous across the whole review. Keep per-comment decorations few; if a comment needs many, consider splitting it.

Communication practices

The standard's communication guide recommends:

Be curious

Don't assume full context. When unsure whether a problem is real, prefer question: over issue: or suggestion:.

Avoid:

**suggestion:** This bug could be solved in the `Main` component. Probably less code.

Prefer:

**question:** Could we solve this in the `Main` component?

It might be a simpler fix.

Leave actionable comments

Make it clear how to resolve the comment. If there's no obvious path forward, say so explicitly.

Combine similar comments

Don't fragment one concern across many tiny comments. Batch related issues into a single comment with a patch or example.

Avoid:

Ten separate **polish:** m_foo → foo comments.

Prefer:

**polish:** Rename all `m_X` variables to `X` — the `m_` prefix isn't used elsewhere in this project.

E.g. `m_foo` → `foo`, `m_bar` → `bar`, `m_quackaroo` → `quackaroo`.

Replace "you" with "we"

"You" places focus on the person; "we" focuses on the task.

Avoid:

**issue:** You should write tests for this.

Prefer:

**todo:** We should write tests for this.

Don't restate the label

The label signals the comment's type, stance, and that something is being said. Don't restate any of that — drop hedging, introductory phrases, fillers, and pleasantries.

Avoid:

**issue:** I think this might be a problem — looking at the code, the function basically swallows exceptions silently. Hope this helps!

Prefer:

**issue:** The function silently swallows exceptions.

Language behavior

  • Labels and decorations are always in English — they're the canonical tokens of the standard and must not be translated.
  • Subject and discussion match the language of the user's message — Russian message, Russian subject and discussion. Same for any other language.

Example (Russian content, English label):

**suggestion (non-blocking):** Стоит вынести валидацию в отдельную функцию.

Сейчас в обработчике смешаны валидация и сохранение — разделение упростит тестирование и переиспользование.

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.