agentleFS
loonylabs-devloonylabs-dev/claude-setup/skills/changelog-summary/SKILL.md

changelog-summary

Skill0 starsChanged 3 months agoMITView on GitHub

Compact summary of the last 7 days of changes, from a technical and an end-user angle, based on git commits and branches. Optional parameter: number of days (e.g. `/changelog-summary 14`).

3 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

3 rules. About 0 are common, 1 is distinctive, and 2 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 2

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Run these commands (replace N with the value you determined):

    line 21
  2. Distinctive:

    Merge commits: only when they close a feature branch (useful context).

    line 153
  3. Very specific:

    Commit messages with fix: → "under the hood".

    line 154

≈ “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

Changelog Summary

First check whether a parameter (number of days) was passed, e.g. /changelog-summary 14.

  • With a parameter: use that value as the period.
  • Without: use 7 days as the default.

State the period you used explicitly in the output.

The changelog itself is written for the product's readers, not for this file. Render every heading, bullet and the social post in the conversation language — the headings below are given in English because this is an instruction; they are not the literal strings to emit. Everything else about language follows language.md.

Step 1: gather the git data

Run these commands (replace N with the value you determined):

# Commits over the last N days (all branches)
git log --all --since="N days ago" --oneline --no-merges --format="%h %s (%an, %ar)"

# Merge commits
git log --all --since="N days ago" --merges --oneline --format="%h %s (%ar)"

# Changed files per commit (for technical depth)
git log --all --since="N days ago" --no-merges --name-only --format="%nCOMMIT: %h %s" | head -200

# Branches active over the last N days
git branch -a --sort=-committerdate | head -20

Step 2: write the summary

Analyse the data and produce the following output:


Output format

## Changes over the last N days ({{DATE_FROM}} – {{DATE_TO}})

---

### ✨ What's new for you

Features and improvements the user can feel, written to excite.
Phrasing: "you can now…", "from today…", "new:".
Only what the user can directly see, touch or experience.
Max. 5-7 points. No technical terms. No jargon.

Shape:
- **[feature name]:** you can now…
- **[feature name]:** from today…

---

### 🔧 Improved under the hood

Performance, stability and technical quality work the user does not see directly but
does notice (faster, more reliable, fewer crashes).
Only things without a visible failure story — no "X was broken".
Phrasing: short and factual, no filenames or API names.
Max. 4-6 points.

Shape:
- **[area]:** what was stabilised or optimised.
- ...

---

### 🐛 Bugfixes

Visible defects that were fixed — wrong behaviour, broken UI, crashes, wrong output.
Only bugs a user would concretely have noticed. No internal refactoring.
Phrasing: what was broken → now fixed. Short, no jargon.
Max. 5-8 points.

Shape:
- **[feature/area]:** [what the problem was] — fixed.
- ...

---

### 🔩 Technical (for developers)

Compact summary grouped by subject area (max. 8-10 points).
Each point: 1-2 sentences, naming the file or module, one concrete technical statement.
No filler. Facts only.

Shape:
- **[area]:** what changed, and why if the commit says.
- **[area]:** ...

---

### 📱 Social post of the week

Pick the **strongest user-facing features** from "✨ What's new for you" — the ones that
excite or surprise most.

**How many bullets:** decide by the quality of the week:
- **5 bullets** when there are many strong, tangible features
- **4 bullets** as a good middle for a solid release
- **3 bullets** when little is user-visible — fewer beats padding

Write a **short social post** that:
- opens with a strong hook (curiosity, surprise, FOMO)
- names the chosen features as short bullets (max. 1 line each, led by an emoji)
- closes with a clear call to action

Tone: direct, energetic, a founder sharing progress — not an advertisement.
Address the reader informally.

**Audience and call to action come from the project, not from this skill.** Take them from
`CLAUDE.md` / `README.md` (product description, USP, domain). If there is nothing usable
there, ask ONCE — do not invent an audience.

Shape (4 bullets):

[hook — 1 line]

✨ [feature 1 — 1 line] 🎙️ [feature 2 — 1 line] 📱 [feature 3 — 1 line] 🚀 [feature 4 — 1 line]

[call to action — product name/domain from the project]


---

### 📊 Activity

- Commits: X (excluding merges)
- Active branches: X
- Most-changed areas: [top 3 directories/modules]

Guidelines

  • "✨ What's new" — genuine user features only: new functions, new content, visible UX improvements. No "bugfix under the hood".
  • "🔧 Under the hood" — bugfixes, performance, stability the user feels indirectly (faster, fewer errors, more reliable). No filenames.
  • "🔩 Technical" — the full technical picture with module references, for developers.
  • Group by subject, not chronologically (e.g. storage, TTS, export, UI, bugfixes).
  • Merge commits only when they close a feature branch (useful context).
  • Commit messages with fix: → "under the hood"; feat: → "what's new"; refactor:/chore: → technical only.
  • Social post: pick the features with the highest wow factor — visible, tangible, surprising. No jargon, no filenames, no architecture topics.

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.