py-changelog
Maintain the release-history.rst changelog for a Python project before publishing a new version. Use when the user wants to prepare release notes, update changelog, or is about to publish/tag a new version.
- License 许可
- MIT
- 最近更新
- 2026-11-07
- Stars
- ★ 0
- Forks
- 0
01 前言 Front Matter
- name
- py-changelog
- description
- Maintain the release-history.rst changelog for a Python project before publishing a new version. Use when the user wants to prepare release notes, update changelog, or is about to publish/tag a new version.
03 SKILL.md 正文
You are maintaining release-history.rst for a Python project. Your job is to keep the target version's section accurate and complete based on git history. You never touch the x.y.z (Backlog) section.
Step 1: Gather State
Run these commands to understand the current state:
# Last published tag
git describe --tags --abbrev=0
# pyproject.toml version (this is the target version to release)
grep '^version' pyproject.toml
Determine target version:
pyproject.toml versionis always the source of truth for the new release target.last tagis the baseline for git history comparison.
Step 2: Pull Git Changes (Feed to AI)
Use {LAST_TAG} = the tag from Step 1.
# Commit intent summary
git log {LAST_TAG}..HEAD --oneline
# Files changed (concise overview)
git diff {LAST_TAG}..HEAD --stat
# Actual Python code changes (for understanding what really changed)
git diff {LAST_TAG}..HEAD -- '*.py'
Read these outputs carefully. They are the ground truth for what changed.
Step 3: Read release-history.rst Format
Read the first 100 lines of release-history.rst:
head -100 release-history.rst
This shows you:
- The RST formatting conventions used in this project
- The section categories (Features and Improvements, Minor Improvements, Bugfixes, Breaking Changes, Miscellaneous)
- The
x.y.z (Backlog)placeholder at the top — never modify this section - The Backlog content — use it as reference only to understand what the user plans, do not copy or move it
Step 4: Determine Action
Case A — Target version section already exists in release-history.rst:
Compare the existing content against what git history reveals:
- Is anything missing that git history shows?
- Is anything written that git history does not support?
- Are the descriptions accurate and clear?
Update the section in-place to be complete and accurate. Do not leave gaps.
Case B — Target version section does not exist:
Insert a new section immediately after the x.y.z (Backlog) block (after its blank line), following this exact RST format:
{VERSION} ({DATE})
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
**Features and Improvements**
- ...
**Minor Improvements**
- ...
**Bugfixes**
- ...
**Miscellaneous**
- ...
Use today's date. Omit any category that has no entries (do not leave empty **Category** headers with no bullets).
Step 5: Write Directly
Edit release-history.rst directly. No need to ask the user for confirmation — the file is git-tracked and can be reverted.
Rules
- Never touch
x.y.z (Backlog)— not its header, not its content, not its blank lines. - Backlog content is reference only. Do not copy bullet points verbatim from Backlog into the version section.
- Base all changelog entries on what git history actually shows, not assumptions.
- Keep descriptions user-facing and concise (what changed and why it matters, not implementation details).
- Follow the RST formatting exactly as seen in existing entries.
- If a category has no entries, omit the entire category header.
04 同仓库更多 Skills
write-pypi-skill
A meta skill that reads the Python library at the current working directory and generates a reusable Agent Skill for it, named after the package, that documents the library's public API and, if present, its command line interface. The user invokes this when they want to turn a local Python library into a Claude skill so future sessions can use the library through its public surface without re-reading the source.
write-agent-skill
Claude Code Agent Skill Author Guide
pypi-chinese_to_english_punctuation
Converts Chinese full-width punctuation into English half-width punctuation and fixes the spacing between Chinese characters and adjacent English words or numbers, while preserving leading indentation so Markdown and reStructuredText code blocks stay intact. Covers commas, periods, colons, semicolons, question and exclamation marks, round brackets, square brackets, angle brackets, curly braces, and double quotes. The single public function is process, which takes text and returns text. A command line tool named c2ep exposes the same behavior, with a text subcommand that reads from stdin or an argument and writes to stdout, and a file subcommand that rewrites a UTF-8 file in place. Use this when normalizing punctuation in mixed Chinese and English documents. Generated for chinese_to_english_punctuation version 0.1.2. Higher installed versions should generally be compatible, lower versions may still work but correctness is not guaranteed.