context-bridge
Synchronize task state and metadata across Claude, Cursor, and Factory Droid sessions. Use when handing off tasks between platforms, sharing plans, or updating external trackers like Linear or Jira.
When & Why to Use This Skill
Context Bridge is a powerful Claude skill designed to synchronize task states, metadata, and execution plans across multiple AI platforms such as Claude, Cursor, and Factory Droid. It eliminates context loss during platform handoffs by maintaining a centralized state, ensuring that AI agents remain aligned with the current project progress and external trackers like Linear, Jira, and GitHub.
Use Cases
- Seamless Platform Handoff: Transition a complex coding session from Claude to Cursor while preserving the exact state of the plan, active artifacts, and next steps.
- Automated Project Management: Automatically update issue statuses and metadata in Linear or GitHub repositories as the AI agent completes tasks across different sessions.
- Artifact Metadata Syncing: Maintain a consistent registry of published artifacts across environments, ensuring that versioning and publishing statuses are always up-to-date.
- Multi-Agent Collaboration: Enable different AI agents to work on the same project by sharing a unified 'memory' file, preventing redundant work and ensuring workflow continuity.
| name | context-bridge |
|---|---|
| version | 1.0.0 |
| description | Synchronize task state and metadata across Claude and Cursor sessions. Use when handing off tasks between platforms, sharing plans, or updating external trackers like Linear or Jira. |
| allowed-tools | linear_read, linear_write, github_read, github_write, slack_write |
Context Bridge - Synchronizes task state and metadata across Claude, Cursor, and Factory Droid sessions. Ensures "memory" is preserved when switching between different AI agents or platforms.
- Synchronizing task state and metadata across platforms
- Handing off tasks between platforms
- Sharing plans and updating external trackers (Linear, Jira)
- Preserving context when switching AI agents
1. **Identify Handoff**: When a user indicates they are switching platforms (e.g., "I'll finish this in Cursor"), invoke this skill.
2. **Persist State**: Save the current plan, active artifacts, and next steps to `.claude/context/state.json`.
3. **Sync Publishing Metadata**:
- Read artifact registry from source platform (e.g., `.claude/context/runtime/runs//artifact-registry.json`)
- Extract publishing metadata (published, published_at, publish_status, publish_targets, publish_attempts)
- Validate metadata against `.claude/schemas/artifact-metadata.schema.json`
- Sync to target platform (e.g., `.cursor/plans/artifacts/`)
- Handle conflicts: prefer most recent `published_at` timestamp
- Update `last_synced` timestamp in metadata
4. **Update Trackers**: If a ticket ID is present, update the external status (Linear/GitHub).
5. **Notify**: Send a summary to the appropriate channel if requested.
- **Linear**: Read/Write issues.
- **GitHub**: Read repo context, update PRs/Issues.
- **Slack**: Send notifications to team channels.
- Always update the central state file before sending notifications.
- Never overwrite existing state without reading it first to preserve history.
- Publishing Metadata Sync: When syncing artifacts across platforms:
- Validate metadata against
.claude/schemas/artifact-metadata.schema.jsonbefore syncing - Handle publishing status conflicts by preferring most recent
published_attimestamp - Preserve
publish_attemptshistory for retry tracking - Update
last_syncedtimestamp to track sync freshness - If metadata conflicts, log conflict resolution in reasoning file
- Validate metadata against