Changelog logo

Changelog

Organization
claude-world
changelog

View and manage the runtime changelog for observability

Overview

Publisherclaude-world
Repositorydirector-mode-lite
Skill namechangelog
Stars
80
Forks
10
Bundled files
Instructions only
LicenseMIT
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • Self-contained

    Everything the model needs lives in the instructions — no extra files to sync.

  • Open source

    Published by claude-world on GitHub. Read the source before you install it.

Installation

Install the Changelog AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/claude-world/director-mode-lite/tree/main/skills/changelog
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/claude-world/director-mode-lite.git /tmp/director-mode-lite
mkdir -p .claude/skills
cp -r /tmp/director-mode-lite/skills/changelog .claude/skills/changelog
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Changelog in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Changelog on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Changelog is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

Changelog Skill

Status: Experimental — uses Claude Code PostToolUse hooks; the hook interface may change in future versions. If hooks don't fire, events can still be logged manually via auto-loop prompts.

Runtime observability changelog that records every change during a development session, so subagents can understand prior context and sessions can be recovered or debugged.


Overview

Automatically, via PostToolUse hooks, the changelog:

  • Records file changes (Write / Edit)
  • Logs test results when tests run (Bash)
  • Records git commits (Bash)
  • Rotates itself when it exceeds 500 lines

Events are stored as append-only JSONL at .director-mode/changelog.jsonl.


Usage

Via /changelog command

bash
/changelog                  # Recent 10 events
/changelog --summary        # Statistics
/changelog --type test      # Filter by event type
/changelog --list-archives  # Show rotated changelogs
/changelog --export log.json
/changelog --archive        # Manually archive current changelog
/changelog --clear          # Clear current changelog

Via Bash

bash
# Last 5 events
tail -n 5 .director-mode/changelog.jsonl | jq '.'

# All file changes
grep '"event_type":"file_' .director-mode/changelog.jsonl

# Count by type
jq -r '.event_type' .director-mode/changelog.jsonl | sort | uniq -c

Event Schema

json
{
  "id": "evt_1705142400_12345",
  "timestamp": "2025-01-13T10:30:00.000Z",
  "event_type": "file_edit",
  "agent": "hook",
  "iteration": 3,
  "summary": "file_edit: Login.tsx",
  "files": ["src/components/Login.tsx"]
}

Event Types

TypeSourceDescription
file_writeHook (Write)File written via the Write tool
file_editHook (Edit)File edited via the Edit tool
test_passHook (Bash)Tests passing
test_failHook (Bash)Tests failing
test_runHook (Bash)Tests ran, result undetermined
commitHook (Bash)Git commit made
session_startauto-loopSession begins
session_endauto-loopSession completes
changelog_rotatedSystemChangelog was rotated

Automatic Logging via Hooks

Configured in .claude/settings.local.json (the shipped template is hooks/settings-hooks.json):

json
{
  "hooks": {
    "PostToolUse": [
      { "matcher": "Write", "hooks": [{ "type": "command", "command": "\"$CLAUDE_PROJECT_DIR\"/.claude/hooks/log-file-change.sh" }] },
      { "matcher": "Edit",  "hooks": [{ "type": "command", "command": "\"$CLAUDE_PROJECT_DIR\"/.claude/hooks/log-file-change.sh" }] },
      { "matcher": "Bash",  "hooks": [{ "type": "command", "command": "\"$CLAUDE_PROJECT_DIR\"/.claude/hooks/log-bash-event.sh" }] }
    ]
  }
}

$CLAUDE_PROJECT_DIR keeps paths portable (resolved at runtime by Claude Code).

ScriptTriggerEvents Logged
log-file-change.shWrite, Editfile_write, file_edit
log-bash-event.shBashtest_pass, test_fail, test_run, commit

Core helpers live in _lib-changelog.sh:

bash
log_event "file_write" "Wrote Login.tsx" "hook" '["src/Login.tsx"]'
archive_changelog   # move current changelog to a timestamped archive
clear_changelog     # empty the current changelog
list_archives       # list rotated changelogs

Rotation & Archiving

Two independent thresholds keep the changelog bounded. They are complementary, not contradictory:

TriggerThresholdWho does itWhen
Session-start archive> 100 lines/auto-loop initAt the start of each auto-loop session, so the new session begins with a fresh log
Runtime auto-rotation> 500 lines (MAX_LINES)_lib-changelog.sh (rotate_if_needed)Continuously, as events are appended within a session

On rotation the current file is moved to changelog.YYYYMMDD_HHMMSS.jsonl, a fresh changelog.jsonl is started, and a changelog_rotated event is logged:

.director-mode/
├── changelog.jsonl                    ← Current
├── changelog.20250113_103000.jsonl    ← Archived
└── changelog.20250112_150000.jsonl    ← Archived

Checkpoint vs Changelog

AspectCheckpointChangelog
Location.auto-loop/checkpoint.json.director-mode/changelog.jsonl
PurposeCurrent state snapshot ("where am I now?")Historical event stream ("how did I get here?")
Used byStop hook (continue/stop decision)Subagents (context)
FormatSingle JSON objectJSONL (append-only)
PersistenceOverwritten each iterationAccumulated, then rotated

Only one auto-loop session runs per project. Starting /auto-loop while a session is in_progress blocks with a prompt to use --resume (continue with the existing checkpoint + changelog) or --force (archive the old session, start fresh).


Subagent Integration

  • code-reviewer — before reviewing, checks recent file changes, the current iteration, and recent test results.
  • debugger — before debugging, checks when errors first appeared, which files changed just before them, and the pattern of test failures.

Installation

These legacy observability hooks ship as an explicit automation option. They are absent from the default zero-hook install. After intentionally running ./install.sh --hooks automation, verify:

bash
ls .claude/hooks/
# _lib-changelog.sh  auto-loop-stop.sh  log-bash-event.sh  log-file-change.sh  pre-tool-validator.sh

cat .claude/settings.local.json | jq '.hooks'

Troubleshooting

Events not logged

  1. Check hooks exist: ls .claude/hooks/*.sh
  2. Check the live hook config: cat .claude/settings.local.json | jq '.hooks' (the shipped template is hooks/settings-hooks.json)
  3. Check scripts are executable: chmod +x .claude/hooks/*.sh

Stale session blocking

bash
cat .auto-loop/checkpoint.json | jq '.status'
/auto-loop --force "New task"

Changelog too large

bash
/changelog --archive   # manual archive
/changelog --clear     # or clear

Frequently asked questions

What does the Changelog AI skill do?

View and manage the runtime changelog for observability

Why use Changelog on TypingMind?

Because you install it once and use it with any model. Changelog is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Changelog in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/claude-world/director-mode-lite/tree/main/skills/changelog. TypingMind reads its SKILL.md and installs it as a skill you can enable per chat.

Which AI models can use Changelog?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Changelog?

As many as you like. As long as a model supports skills, you can use Changelog with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Changelog AI skill free?

Yes. It is published on GitHub by claude-world under the MIT license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇