Checkpoint logo

Checkpoint

Community
jmagly
checkpoint

Create, list, or recover mid-workflow checkpoints so interrupted work resumes from a known-good position

Overview

Publisherjmagly
Repositoryaiwg
Skill namecheckpoint
Stars
225
Forks
34
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 jmagly on GitHub. Read the source before you install it.

Installation

Install the Checkpoint 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/jmagly/aiwg/tree/main/agentic/code/addons/aiwg-utils/skills/checkpoint
  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/jmagly/aiwg.git /tmp/aiwg
mkdir -p .claude/skills
cp -r /tmp/aiwg/agentic/code/addons/aiwg-utils/skills/checkpoint .claude/skills/checkpoint
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Checkpoint 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 Checkpoint 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 Checkpoint 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.

Checkpoint

You create, list, or recover lightweight mid-workflow checkpoints that allow a crashed or interrupted workflow to continue from a known-good position rather than restart from scratch.

Triggers

Alternate expressions and non-obvious activations (primary phrases are matched automatically from the skill description):

  • "save my place" → create checkpoint at current position
  • "mark progress" → create checkpoint
  • "pick up where we left off" → recover from most recent checkpoint
  • "what checkpoints exist" → list checkpoints
  • "resume from checkpoint" → recover from specified or latest checkpoint

Trigger Patterns Reference

PatternExampleAction
Create (named)"checkpoint this as iteration-3-complete"Run aiwg checkpoint create --name iteration-3-complete
Create (anonymous)"create a checkpoint"Run aiwg checkpoint create
List"show available checkpoints"Run aiwg checkpoint list
Recover (latest)"recover from the last checkpoint"Run aiwg checkpoint recover
Recover (specific)"recover checkpoint ckpt_7b2f1a"Run aiwg checkpoint recover ckpt_7b2f1a

Behavior

When triggered:

  1. Extract intent:

    • Which subcommand: create, list, or recover?
    • Is a name or checkpoint ID provided?
    • For recover, should it use the most recent checkpoint or a specific one?
  2. Run the appropriate subcommand:

    bash
    # Create — anonymous
    aiwg checkpoint create
    
    # Create — named
    aiwg checkpoint create --name <name>
    
    # List all checkpoints
    aiwg checkpoint list
    
    # Recover — most recent checkpoint
    aiwg checkpoint recover
    
    # Recover — specific checkpoint by ID or name
    aiwg checkpoint recover <id>
  3. Checkpoints vs. snapshots: Checkpoints are lightweight — they record the current workflow phase and step plus any in-progress artifact paths, not full artifact checksums. Use snapshots for complete point-in-time captures; use checkpoints for crash recovery during active workflows.

    CheckpointSnapshot
    PurposeCrash recoveryReproducibility
    SizeSmall (phase + step + artifact refs)Full (versions + checksums)
    Created byManual or agent loop (auto)Manual only
    Use forResume interrupted runReplay from known state
  4. Automatic creation: Agent loops create checkpoints automatically between iterations. Manual create supplements this for critical workflow milestones.

  5. Storage: Checkpoints are stored in .aiwg/checkpoints/ as lightweight JSON files.

  6. Report the result — on create, confirm the checkpoint ID and position; on recover, confirm which step the workflow will resume from.

Examples

Example 1: Create a named checkpoint

User: "Create a checkpoint after iteration 3 completes"

Extraction: Create subcommand, name iteration-3-complete

Action:

bash
aiwg checkpoint create --name iteration-3-complete

Response: "Checkpoint iteration-3-complete created (ID: ckpt_7b2f1a). Recorded position: Construction phase, iteration 3, step 12/15. In-progress artifacts: .aiwg/working/tests/integration-plan-draft.md. Stored at .aiwg/checkpoints/ckpt_7b2f1a.json."

Example 2: List checkpoints

User: "What checkpoints do we have?"

Extraction: List subcommand

Action:

bash
aiwg checkpoint list

Response:

ID            Name                     Date                  Phase / Step
ckpt_3a8c12   (auto)                   2026-03-28 10:04:51   Construction / iter 1, step 8
ckpt_9f20de   (auto)                   2026-03-29 14:17:33   Construction / iter 2, step 15
ckpt_7b2f1a   iteration-3-complete     2026-03-30 09:55:02   Construction / iter 3, step 12

Example 3: Recover from the latest checkpoint

User: "The workflow crashed — recover from the last checkpoint"

Extraction: Recover subcommand, use most recent

Action:

bash
aiwg checkpoint recover

Response: "Recovering from most recent checkpoint iteration-3-complete (ckpt_7b2f1a). Resuming Construction phase, iteration 3, step 12/15. In-progress artifact .aiwg/working/tests/integration-plan-draft.md reloaded. Ready to continue."

Example 4: Recover from a specific checkpoint

User: "Recover from checkpoint ckpt_9f20de"

Extraction: Recover subcommand, ID ckpt_9f20de

Action:

bash
aiwg checkpoint recover ckpt_9f20de

Response: "Recovering from checkpoint ckpt_9f20de (auto, 2026-03-29 14:17:33). Resuming Construction phase, iteration 2, step 15/15. No in-progress artifacts. Ready to continue from this position."

Clarification Prompts

If the user's intent is ambiguous:

  • "Should I create a new checkpoint here, list existing checkpoints, or recover from one?"
  • "Which checkpoint would you like to recover from — the most recent, or a specific one? Run aiwg checkpoint list to see what's available."

References

  • @$AIWG_ROOT/src/cli/handlers/subcommands.ts — Checkpoint command handler
  • @$AIWG_ROOT/docs/cli-reference.md — CLI reference
  • @$AIWG_ROOT/agentic/code/addons/aiwg-utils/skills/snapshot/SKILL.md — Full workflow snapshots (compare to checkpoints)

Frequently asked questions

What does the Checkpoint AI skill do?

Create, list, or recover mid-workflow checkpoints so interrupted work resumes from a known-good position

Why use Checkpoint on TypingMind?

Because you install it once and use it with any model. Checkpoint 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 Checkpoint in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/jmagly/aiwg/tree/main/agentic/code/addons/aiwg-utils/skills/checkpoint. TypingMind reads its SKILL.md and installs it as a skill you can enable per chat.

Which AI models can use Checkpoint?

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 Checkpoint?

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

Is the Checkpoint AI skill free?

Yes. It is published on GitHub by jmagly 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 👇