Checkpoint Workflow System

VerifiedSafe

This skill provides a workflow checkpoint system that tracks progress phases (setup, research, implement, review, finalize) and logs actions and commits. It uses CLI commands via `pnpm checkpoint workflow` to create, find, update, and resume workflows. Helpful for maintaining context across sessions and enabling resume after agent context compaction.

Sby Skills Guide Bot
DevelopmentIntermediate
6/2/2026
Claude Code
#checkpoint#workflow#resume#cli#issue-tracking

Recommended for

Automated summary

This skill manages a checkpoint system for issue resolution workflows, enabling seamless resumption after context compaction.

Strengths

  • Structures work into clear phases (setup, research, implement, review, finalize)
  • Allows resuming exactly where you left off
  • Logs actions and commits for full traceability

Limitations

  • Requires a local SQLite database that may be lost
  • CLI commands must be executed precisely without shell variables
  • Only works for workflows managed through this specific system
When to use it

Use when working on a complex issue that spans multiple sessions or requires resumption after context clearance.

When not to use it

Avoid for simple or one-off tasks that do not require state tracking.

Security analysis

Safe
Quality score90/100

The skill describes internal CLI commands for a workflow checkpoint system. No external scripts, destructive operations, or exfiltration are instructed. It does not introduce any realistic security risk.

No concerns found

How this analysis is made

Examples

Create a new workflow for an issue
Create a new checkpoint workflow for issue 42 on branch 'fix/42-null-pointer'.
Find an existing workflow by issue number
Check if there is an existing workflow for issue 15 and show its status.
Log a commit after creating it
Log the latest commit (SHA from git rev-parse HEAD) with message 'fix: resolve null pointer' into the current workflow.

Checkpoint Workflow Skill

Reference for agents working with the workflow checkpoint system.

Purpose

The checkpoint system enables workflow resume after context compaction. Each /work-on-issue session creates a workflow that tracks:

  • Current phase (setup → research → implement → review → finalize)
  • Actions taken (plan created, PR opened, etc.)
  • Commits made during the workflow

Database Location

Checkpoint state is stored in .claude/execution-state.db (gitignored).

CLI Commands

All commands use pnpm checkpoint workflow <action>.

Create Workflow

pnpm checkpoint workflow create <issue_number> "<branch_name>"

Output: JSON with id, issue_number, branch_name, status, current_phase

Example:

pnpm checkpoint workflow create 12 "feat/12-add-parser"

Find Workflow by Issue

pnpm checkpoint workflow find <issue_number>

Output: JSON workflow or null

Use case: Check if workflow exists before creating a new one.

Get Workflow Summary

pnpm checkpoint workflow get <workflow_id>

Output: JSON with workflow, actions, commits

Use case: Full context for resuming work.

List Workflows

pnpm checkpoint workflow list [--status=running|completed|failed] [--limit=N]

Output: JSON array of workflows

Set Phase

pnpm checkpoint workflow set-phase <workflow_id> <phase>

Phases: setup, research, implement, review, finalize

Set Status

pnpm checkpoint workflow set-status <workflow_id> <status>

Statuses: running, paused, completed, failed

Log Action

pnpm checkpoint workflow log-action <workflow_id> <action_type> <status> [details]

Status values: success, failed, skipped

Common action types:

  • workflow_started
  • dev_plan_created
  • implementation_complete
  • pr_created

Log Commit

pnpm checkpoint workflow log-commit <workflow_id> <sha> "<message>"

CRITICAL: Commit logging pattern

Always log commits in two separate commands:

# 1. Get SHA (separate command)
git rev-parse HEAD

# 2. Log to checkpoint (use literal SHA, not variable)
pnpm checkpoint workflow log-commit "abc123" "a1b2c3d" "feat: add feature"

NEVER combine with && or use shell variables. This prevents errors if the git command fails.

Delete Workflow

pnpm checkpoint workflow delete <workflow_id>

Integration Patterns

At Workflow Start (setup-agent)

# Check for existing workflow
pnpm checkpoint workflow find {issue_number}

# If found with status=running, offer to resume
# If not found, create new workflow
pnpm checkpoint workflow create {issue_number} "{branch_name}"

After Plan Creation (issue-researcher)

pnpm checkpoint workflow set-phase "{workflow_id}" research
pnpm checkpoint workflow log-action "{workflow_id}" "dev_plan_created" success

After Each Commit (atomic-developer)

# Get SHA first
git rev-parse HEAD
# Then log (with actual SHA value)
pnpm checkpoint workflow log-commit "{workflow_id}" "{sha}" "{message}"

After PR Creation (finalize-agent)

pnpm checkpoint workflow log-action "{workflow_id}" "pr_created" success
pnpm checkpoint workflow set-status "{workflow_id}" completed

Resume Flow

When starting /work-on-issue:

  1. Check for existing workflow: pnpm checkpoint workflow find {issue_number}
  2. If found with status=running:
    • Show current phase and recent actions
    • Ask: "Resume from {phase}?" or "Start fresh?"
  3. If resuming, skip to the saved phase
  4. If starting fresh, delete old workflow and create new one

Error Handling

If a checkpoint command fails:

  1. Log the error
  2. Continue with the workflow (checkpoints are advisory, not blocking)
  3. The work can still be done, just without resume capability

From the file to a first result

The download contains exactly the content displayed on this page. This copy may differ from the latest version in the source repository.

View the source and supporting files

This download contains SKILL.md only. Any scripts, references, templates and dependencies are not included. Get the required files from the source, preserving their directory structure, before continuing.

1. Download the file

Download SKILL.md

From the root of a test project, save the file as .skill-downloads/checkpoint-workflow-system-mm5t8ada/SKILL.md, or use the command below. This folder is for inspecting the skill before installation.

Command for macOS/Linux or WSL: requires a POSIX shell and curl. On Windows without WSL, use the download button and create the folder manually.

mkdir -p '.skill-downloads/checkpoint-workflow-system-mm5t8ada' && curl --fail --show-error --location 'https://skills-guides.com/api/skills/checkpoint-workflow-system-mm5t8ada/download' --output '.skill-downloads/checkpoint-workflow-system-mm5t8ada/SKILL.md.part' && mv '.skill-downloads/checkpoint-workflow-system-mm5t8ada/SKILL.md.part' '.skill-downloads/checkpoint-workflow-system-mm5t8ada/SKILL.md'

2. Check prerequisites

Read SKILL.md before using it. If it references scripts or external services, inspect them and prepare the required access in your test environment.

No recognized prerequisites section in this file. This does not mean it is self-contained: libraries, tools and access requirements may be mentioned elsewhere in the content or source repository.

The platforms on this page are catalog indications. The paths below follow tool documentation; they do not prove this skill has been run in those tools. No minimum tool version is confirmed for this listing.

3. Load the skill

This file does not provide a usable name/description header for this native procedure. It remains available to download and inspect; use the manual method below.

Manual method or another tool: attach SKILL.md to a conversation, or open it in an agent that can read files. Include any required supporting resources. No universal folder is recognized by every tool.

4. Try a first task

Use a short, synthetic input: an API endpoint, a small document or an interface brief. In a new conversation, attach the file or give the agent its path, then adapt this request:

Read .skill-downloads/checkpoint-workflow-system-mm5t8ada/SKILL.md and apply its instructions to this example: [describe a test input]. Identify any missing files, dependencies or access. Produce the result requested by the skill without inventing data or running unreviewed scripts.

Expected result: output matching the skill instructions, or a precise list of missing prerequisites. A read confirmation alone does not validate a first use. If an external resource is unavailable, that path remains untested.

Related skills