/ase-meta-changelog

name: ase-meta-changelog
argument-hint: "[--help|-h]"
description: >
    Update ChangeLog entries in CHANGELOG.md files
user-invocable: true
disable-model-invocation: false
model: sonnet
effort: high
allowed-tools:
    - "Bash(git log *)"
    - "Bash(git diff *)"
    - "Bash(git show *)"
    - "Bash(git tag --list *)"
    - "Read"
    - "Write"
    - "Edit"

@${CLAUDE_SKILL_DIR}/../../meta/ase-control.md @${CLAUDE_SKILL_DIR}/../../meta/ase-skill.md @${CLAUDE_SKILL_DIR}/../../meta/ase-getopt.md

<purpose name="ase-meta-changelog"> Update ChangeLog Entries </purpose>

<expand name="getopt" arg1="ase-meta-changelog"> $ARGUMENTS </expand>

<objective> Help to complete, consolidate and sort *ChangeLog* entries of the most recent *ChangeLog* section, based on underlying *Git* commits and staged changes. </objective>

Format ------

The *ChangeLog* file is a Markdown formatted file named `CHANGELOG.md`, and contains sections with headers in the style `N.M.K (YYYY-MM-DD)`.

Each *ChangeLog* entry is always formatted as...

`<change-type/> [<artifact-kind/>]: <summary/>`

...where the <change-type/> is one of the following tags and their usual related changes:

• `FEATURE`: new functionality or configuration
• `IMPROVEMENT`: improved functionality or configuration
• `BUGFIX`: corrected functionality or configuration
• `UPDATE`: updated functionality or configuration
• `CLEANUP`: cleaned up functionality or configuration
• `REFACTOR`: refactored functionality or configuration

The <artifact-kind/> is one or more of the following *artifact* tags, classifying which kind of artifact the change primarily touches. If multiple artifact kinds apply, comma-separate them (e.g. `[arch, code]`):

• `spec`: requirement/specification artifacts
• `arch`: architecture/design artifacts
• `code`: source code artifacts
• `docs`: documentation artifacts
• `infr`: infrastructure/build/tooling artifacts
• `othr`: any other artifacts

The <summary/> is not longer than about 60-80 characters. The *ChangeLog* entries for a single product release version are also always grouped and sorted according to the above <change-type/> list.

Processing ----------

<flow>

1. <step id="STEP 1: Locate and read ChangeLog entries">

The *ChangeLog* file `CHANGELOG.md` is located in the *current* directory or one of the *parent* directories of the current project. Locate and read this file. Store its relative path in <filename/>.

You *MUST* *NOT* output anything, except the result with the following <template/>:

<template> <ase-tpl-bullet-normal/> **CHANGELOG FILE:** `<filename/>` </template>

</step>

2. <step id="STEP 2: Determine artifact changes">

You *MUST* *NOT* output anything, except introduce the current operation with an output based on the following <template/>:

<template> <ase-tpl-bullet-normal/> **DETERMINE ARTIFACT CHANGES:** </template>

To update the entries of the most recent *ChangeLog* section, consult the Git *commits* plus the currently already staged changes in the Git *index*, but *ignore* the Git *stash* and still unstaged changes.

For finding the corresponding Git *commits*, first determine the correct baseline Git tag. Take the `N.M.K` from the *first* (most recent) level-2 header in the *ChangeLog* file and check whether a corresponding Git tag already exists with the command `git tag --list N.M.K`. If this command *does* produce output, the first section is already released/tagged, so use the `N.M.K` from the *first* level-2 header as the baseline tag and *insert* a new, still-unreleased level-2 section (next patch version, current date) *above* it, which then becomes the *first* (most recent) section all subsequent steps operate on - the already released section *MUST* *NOT* be modified. If this command produces *no* output, the first section is still in-progress/untagged, so use the `N.M.K` from the *second* level-2 header as the baseline tag instead. If *no* second level-2 header exists (very first, still untagged release), skip the baseline tag and check *all* Git commits with the command `git log HEAD --numstat --pretty=format:'%h: %s'` instead. Then check all Git commits between `HEAD` and this baseline tag with the command `git log N.M.K..HEAD --numstat --pretty=format:'%h: %s'`.

For finding the corresponding staged Git *changes* in the Git *index*, use the command `git diff --cached --numstat`, but silently skip already existing changes to the `CHANGELOG.md` file itself. If still no corresponding *ChangeLog* entry exists for these staged Git *changes*, derive a meaningful one from a `git diff --cached` command.

</step>

3. <step id="STEP 3: Complete ChangeLog entries">

You *MUST* *NOT* output anything, except introduce the current operation with an output based on the following <template/>:

<template> <ase-tpl-bullet-normal/> **COMPLETE ENTRIES:** </template>

Without immediately modifying the `CHANGELOG.md` file, *complete* the entries in the first (most recent) section only, by adding the corresponding (most recent) Git *commits* and *staged* changes only.

For each Git commit, reduce the Git commit messages to a single short <summary/> sentence, not longer than 60-80 characters.

For each entry, also determine the <artifact-kind/> *artifact kind* tag(s) from the paths of the changed files. To classify a changed file to its artifact class, call the `ase_artifact_list(kind: [ ... ])` tool of the `ase` MCP server *once*, passing the `kind` tokens (`spec`, `arch`, `code`, `docs`, `infr`, `othr`), and read the returned `artifacts` array of `{ kind, files }` objects to match each changed file to its kind. Use the matched lower-cased k