Skip to content
Back to skills

2530 03 Auto Checkpoint Hooks 9b1442c1

ASecurity

> **Note:** Checkpoint and iteration MCP tools (`checkpoint_create`, `checkpoint_resume`, `iteration_start`, `iteration_validate`, `iteration_next`, `iteration_complete`) have been removed. Agents now self-manage their iteration loops, and checkpoint/recovery functionality is handled via git commits and Memory Copilot. This document is retained for historical reference.

  • 9 stars
  • 0 votes
  • 0 copies
  • 1 view
  • Added May 31, 2026
documentationbashgitdocumentation

Works with

  • cli
  • mcp

Security analysis

A100/100

Pro scans all 2 files and shows the line behind each finding

Scanned May 31, 2026

npx -y skills add tools-only/X-Skills --skill 2530-03-auto-checkpoint-hooks_9b1442c1 --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of 2530 03 Auto Checkpoint Hooks 9b1442c1?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for 2530 03 Auto Checkpoint Hooks 9b1442c1
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/tools-only-2530-03-auto-checkpoint-hooks-9b1442c1/badge)](https://www.skillsdirectory.com/skills/tools-only-2530-03-auto-checkpoint-hooks-9b1442c1)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
# Auto-Checkpoint Hooks (Legacy)

> **Note:** Checkpoint and iteration MCP tools (`checkpoint_create`, `checkpoint_resume`, `iteration_start`, `iteration_validate`, `iteration_next`, `iteration_complete`) have been removed. Agents now self-manage their iteration loops, and checkpoint/recovery functionality is handled via git commits and Memory Copilot. This document is retained for historical reference.

## Overview

The auto-checkpoint system previously created recovery points during iteration loops automatically. With the migration to the `tc` CLI, checkpoint and iteration tools have been removed. Agents self-manage their iteration loops using standard tooling (run tests, check results, iterate).

## What Changed

| Old MCP Tool | Status |
|---|---|
| `checkpoint_create(...)` | Removed |
| `checkpoint_resume(...)` | Removed |
| `iteration_start(...)` | Removed (agents self-manage) |
| `iteration_validate(...)` | Removed (agents self-manage) |
| `iteration_next(...)` | Removed (agents self-manage) |
| `iteration_complete(...)` | Removed (agents self-manage) |

## Current Approach

Recovery and checkpointing are now handled through:

- **Git commits** -- Auto-commit on task completion creates recovery points
- **Memory Copilot** -- `initiative_update` saves session state for `/continue`
- **Task Copilot CLI** -- `tc task get <id> --json` retrieves task state
- **Work products** -- `tc wp store --task <id> --type <t> --title "..." --content "..." --json` persists deliverables

## Agent Iteration Pattern (Current)

Agents manage their own iteration loops without dedicated MCP tools:

```markdown
FOR EACH attempt (up to max retries):
  # Do work
  - Read files
  - Make changes
  - Run tests via Bash

  # Check results
  IF tests pass AND lint clean:
    Store work product: tc wp store --task <id> ...
    Update task: tc task update <id> --status completed --json
    BREAK
  ELSE:
    Analyze failures, refine approach, try again
```

## Blocked Stream Compaction

When an agent encounters a blocker, it should:
1. Store findings as a work product: `tc wp store --task <id> --type other --title "Blocked: ..." --content "..." --json`
2. Update task status: `tc task update <id> --status blocked --json`

## Related Documentation

- [Working Protocol](../30-operations/01-working-protocol.md) - Agent-First Protocol
- [Agent Guide](../30-operations/03-agent-guide.md) - Agent configuration

Files in this skill

  • README.md1.1 KB
  • skill.md2.4 KB

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.