crash-recovery
About
This skill recovers development sessions after crashes or context loss by reconstructing state and creating handoff notes. It activates automatically following crashes, terminations, or risky operations to prevent data loss. Key features include checkpoint creation, state reconstruction, and seamless recovery with git context integration.
Quick Install
Claude Code
Recommended/plugin add https://github.com/majiayu000/claude-skill-registrygit clone https://github.com/majiayu000/claude-skill-registry.git ~/.claude/skills/crash-recoveryCopy and paste this command in Claude Code to install this skill
Documentation
Crash Recovery Skill
Session continuity and post-crash state reconstruction specialist. Ensures seamless recovery after IDE crashes, context resets, or unexpected terminations.
When This Skill Activates
- After crashes or unexpected terminations
- Context loss or session reset
- Before risky operations (proactive checkpoint)
- When resuming work after long break
- Recovering in-progress tasks
- Reconstructing session timeline
Purpose
CRASH_RECOVERY_SPECIALIST ensures session continuity by:
- Creating periodic checkpoints before risky operations
- Reconstructing session state from git and scratchpad artifacts
- Generating handoff notes for seamless recovery
- Implementing "Prior You / Current You" protocol
Key Insight: Future you depends on current you leaving breadcrumbs.
Reports To
- SYNTHESIZER (Special Staff - Session Continuity)
- May coordinate with HISTORIAN for session documentation
- Provides recovery instructions to ORCHESTRATOR
Agent Identity
Loads: /home/user/Autonomous-Assignment-Program-Manager/.claude/Agents/CRASH_RECOVERY_SPECIALIST.md
Key Workflows
Workflow 1: Checkpoint Creation
INPUT: Current task state before risky operation
OUTPUT: Checkpoint file in .claude/Scratchpad/
1. Snapshot git state (branch, status, recent commits)
2. Capture scratchpad state
3. Document in-progress tasks
4. Create recovery breadcrumbs
5. Save to checkpoint file
Use before: git operations, large refactors, system updates
Workflow 2: Post-Crash Reconstruction
INPUT: Available artifacts (git status, scratchpad files)
OUTPUT: Session state reconstruction
1. Parse git status for uncommitted work
2. Read scratchpad files for context
3. Identify in-progress tasks from history
4. Reconstruct session timeline
5. Cross-reference multiple sources
6. Generate handoff notes
Workflow 3: Handoff Note Generation
INPUT: Recovered state
OUTPUT: Resume notes for next session
1. Summarize recovered state
2. List incomplete tasks
3. Provide recommended first action
4. Flag uncertainties
5. Document what couldn't be determined
"Prior You / Current You" Protocol
Incremental work habits that enable recovery:
- Commit incrementally (~30 min intervals)
- Write to disk, not just memory
- Leave breadcrumbs in scratchpad
- Document decisions and context
This protocol makes recovery possible even without explicit checkpoints.
Output Formats
Checkpoint File
# Recovery Checkpoint: [timestamp]
## Session State
- Branch: [name]
- Last Commit: [hash] [message]
- Uncommitted: [list files]
## Current Task
[What's being worked on]
## Progress
- Completed: [list]
- In Progress: [item]
- Not Started: [list]
## Recent Decisions
[Key choices and rationale]
## Resume Point
[Where to start when recovering]
Recovery Handoff
# Session Recovery: [timestamp]
## Recommended First Action
[Single clear next step]
## Recovered State
- Branch: [name]
- In-progress work: [description]
- Last completed: [task]
## Timeline Reconstruction
[What happened in session]
## Uncertainties
[What couldn't be determined]
## Full Context
[Detailed reconstruction]
Integration with Other Skills
With historian
Coordination: Recovery findings may inform session narratives
1. CRASH_RECOVERY_SPECIALIST reconstructs session
2. If session was significant, findings passed to HISTORIAN
3. HISTORIAN creates narrative documentation
With session-end
Coordination: Crash recovery complements normal session handoff
Normal flow: session-end creates structured handoff
Crash flow: crash-recovery reconstructs what session-end would have captured
Aliases
/recover- Quick recovery invocation/restore- Explicit state restoration
Usage Examples
Example 1: Create Checkpoint Before Risky Operation
Use the crash-recovery skill to create a checkpoint before attempting
a complex git rebase.
## Current State
- Branch: feature/swap-improvements
- Task: Rebasing 15 commits to clean up history
- Progress: All tests passing, ready to rebase
- Recent decisions: Decided to squash related commits
## Output
Save checkpoint to .claude/Scratchpad/checkpoint_[timestamp].md
Example 2: Recover After Crash
Use the crash-recovery skill to recover session state after crash.
## Available Artifacts
- Git branch: feature/resilience-dashboard
- Git status:
On branch feature/resilience-dashboard
Changes not staged for commit:
modified: backend/app/resilience/dashboard.py
modified: frontend/components/Dashboard.tsx
Untracked files:
backend/tests/test_dashboard.py
- Scratchpad files:
.claude/Scratchpad/ORCHESTRATOR_ADVISOR_NOTES.md
.claude/Scratchpad/TODO.md
- Last known task: Implementing resilience dashboard endpoints
## Request
1. Reconstruct session timeline
2. Identify in-progress tasks
3. Generate handoff notes
4. Recommend first action to resume work
Example 3: Routine Recovery at Session Start
Use the crash-recovery skill to review what was happening last session.
Context: Starting new session, want to understand where we left off.
Available:
- Git log (last 10 commits)
- Scratchpad files list
- Current branch: main
Return: Brief summary of last session's work and recommended first step.
Common Failure Modes
| Failure Mode | Symptom | Recovery |
|---|---|---|
| Incomplete Checkpoints | Missing key context | Cross-reference multiple artifact sources; ask user |
| Stale Scratchpad | Scratchpad reflects old state | Use git commits as ground truth; flag discrepancies |
| Uncommitted Work Loss | Work vanished after crash | Check IDE autosave, temp files, .git/ objects |
| Context Ambiguity | Multiple interpretations | Present options to user; avoid assumptions |
| Handoff Overwhelm | Too detailed, can't find resume point | Lead with "Recommended First Action" |
Quality Checklist
Before completing recovery:
- Git state captured (branch, status, commits)
- Scratchpad files reviewed
- In-progress tasks identified
- Timeline reconstructed
- Uncertainties flagged
- Recommended first action provided
- Handoff notes clear and actionable
- Cross-referenced multiple sources
Recovery Artifact Locations
Always Check These Files
.claude/Scratchpad/ORCHESTRATOR_ADVISOR_NOTES.md- Session history.claude/Scratchpad/*.md- Session artifactsCHANGELOG.md- Recent committed changes- Git log (last 10 commits) - Timeline reconstruction
- Git status - Uncommitted work
Checkpoint Storage
Checkpoints saved to: .claude/Scratchpad/checkpoint_[timestamp].md
Context Isolation Awareness
When delegating to CRASH_RECOVERY_SPECIALIST:
For Checkpoint Creation:
- Current task description
- In-progress work status
- Recent decisions made
- Risky operation about to be performed
For Post-Crash Recovery:
- Git branch and full status output
- List of scratchpad files available
- Last known task (if available)
- Approximate time of crash
Recovery scope:
- What needs to be recovered
- Who will resume the work
- Urgency level
Success Metrics
- Recovery completes < 10 minutes
- 95%+ in-progress tasks identified
- Zero data loss from uncommitted work
- Handoff enables seamless resume
Best Practices
Before Risky Operations
Always create checkpoint before:
- Major git operations (rebase, merge, reset)
- Large refactoring
- Database migrations
- System configuration changes
- Complex multi-step procedures
Continuous Recovery Preparation
Implement "Prior You / Current You" protocol:
- Commit every 30 minutes with descriptive messages
- Update scratchpad with decisions
- Write notes explaining "why" not just "what"
- Keep TODO list current
After Crash
- Don't panic
- Run crash-recovery skill immediately
- Review recovered state carefully
- Verify git status matches expectations
- Check for uncommitted work before continuing
- Document what was lost (if anything)
References
- Session history:
.claude/Scratchpad/ORCHESTRATOR_ADVISOR_NOTES.md - Recent changes:
CHANGELOG.md - Checkpoint location:
.claude/Scratchpad/checkpoint_*.md - Recovery patterns:
.claude/dontreadme/synthesis/PATTERNS.md
"Future you depends on current you leaving breadcrumbs."
GitHub Repository
Related Skills
algorithmic-art
MetaThis Claude Skill creates original algorithmic art using p5.js with seeded randomness and interactive parameters. It generates .md files for algorithmic philosophies, plus .html and .js files for interactive generative art implementations. Use it when developers need to create flow fields, particle systems, or other computational art while avoiding copyright issues.
subagent-driven-development
DevelopmentThis skill executes implementation plans by dispatching a fresh subagent for each independent task, with code review between tasks. It enables fast iteration while maintaining quality gates through this review process. Use it when working on mostly independent tasks within the same session to ensure continuous progress with built-in quality checks.
executing-plans
DesignUse the executing-plans skill when you have a complete implementation plan to execute in controlled batches with review checkpoints. It loads and critically reviews the plan, then executes tasks in small batches (default 3 tasks) while reporting progress between each batch for architect review. This ensures systematic implementation with built-in quality control checkpoints.
cost-optimization
OtherThis Claude Skill helps developers optimize cloud costs through resource rightsizing, tagging strategies, and spending analysis. It provides a framework for reducing cloud expenses and implementing cost governance across AWS, Azure, and GCP. Use it when you need to analyze infrastructure costs, right-size resources, or meet budget constraints.
