Skip to content

Latest commit

 

History

History
210 lines (155 loc) · 10.5 KB

File metadata and controls

210 lines (155 loc) · 10.5 KB

Changelog

All notable changes to Mission Control will be documented in this file.

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.


1.4.0 - 2026-03-03

Added

  • Multi-Agent Workflow Pipeline — Full task lifecycle now supports staged orchestration: planning → inbox → assigned → in_progress → testing → review → verification → done.
  • Core Agent Bootstrap — New workspaces can auto-bootstrap a 4-agent core team: Builder (🛠️), Tester (🧪), Reviewer (🔍), and Learner (📚).
  • Workflow Engine Coordination — Added queue-aware review draining (drainQueue()), automatic role-based stage handoffs, and fail-loopback routing.
  • Learner Knowledge Loop — Learner notifications on stage transitions plus knowledge injection into future dispatch messages.
  • New API Routes
    • POST /api/tasks/[id]/fail
    • GET /api/tasks/[id]/roles
    • POST /api/workspaces/[id]/knowledge
    • GET /api/workspaces/[id]/workflows

Changed

  • Strict template defaults — Strict workflow is now default, with review as queue stage and verification owned by the reviewer role.
  • Workspace initialization — New workspaces can clone workflow templates and bootstrap core agents automatically.
  • Project branding/docs — Updated project branding to Autensa (formerly Mission Control) and added explicit privacy-first statement in docs.

Fixed

  • Role mismatch — Fixed strict template verification role (verifierreviewer).
  • Review queue bypass — Fixed auto-advance behavior that could skip proper review queue flow.
  • Dispatch status transition — Fixed dispatch route using hardcoded done; now uses computed next workflow status.
  • Assigned-status resolution — Fixed mapping so assigned resolves to builder stage dispatch correctly.
  • Task template assignment — Fixed task creation path so default workflow template is attached automatically.
  • Learner role assignment — Fixed missing task_roles learner assignment so the learner receives transition events.

Migration

  • Migration 013: Fresh Start — Resets runtime task/agent/event data, sets Strict as default workflow template, and bootstraps core agents for the default workspace.

1.3.0 - 2026-03-02

Added

  • Agent Activity Dashboard — Dedicated page for monitoring agent work with mobile card layout. (#48 — thanks @pkgaiassistant-droid!)
  • Remote Model Discovery — Discover AI models from OpenClaw Gateway via MODEL_DISCOVERY=true env var. (#43 — thanks @davetha!)
  • Proxy Troubleshooting — Added docs for users behind HTTP proxies experiencing 502 errors on agent callbacks.

Fixed

  • Force-Dynamic API Routes — All API routes now use force-dynamic to prevent stale cached responses. (#43)
  • Null Agent Assignmentassigned_agent_id can now be null in task creation schema. (#38 — thanks @JamesCao2048!)
  • Dispatch Spec Forwarding — Planning spec and agent instructions now forwarded in dispatch messages. (#51)
  • Dispatch Failure Recovery — Tasks stuck in pending_dispatch auto-reset to planning status. (#52)

1.2.0 - 2026-02-19

Added

  • Gateway Agent Discovery — Import existing agents from your OpenClaw Gateway into Mission Control. New "Import from Gateway" button in the agent sidebar opens a discovery modal that lists all Gateway agents, shows which are already imported, and lets you bulk-import with one click. Imported agents display a GW badge for provenance tracking. (#22 — thanks @markphelps!)
  • Docker Support — Production-ready multi-stage Dockerfile, docker-compose.yml with persistent volumes, and .dockerignore. Runs as non-root, uses dumb-init for signal handling, includes health checks. (#21 — thanks @muneale!)
  • Agent Protocol Conventions — Added PROGRESS_UPDATE and BLOCKED message formats to the Agent Protocol docs to prevent agent stalling. (#24 — thanks @nice-and-precise!)

Fixed

  • Planning Flow Improvements — Refactored polling to prevent stale state issues, fixed "Other" free-text option (case mismatch bug), made due_date nullable, increased planning timeout to 90s for larger models, auto-start polling on page load. (#26 — thanks @JamesTsetsekas!)
  • WebSocket RPC Deduplication Bug — The event deduplication cache was silently dropping repeated RPC responses with the same payload hash, causing request timeouts. RPC responses now bypass dedup entirely.
  • Next.js Response Caching — Dynamic API routes that query live state (e.g., agent discovery) now use force-dynamic to prevent stale cached responses.

1.1.0 - 2026-02-16

🔒 Security

  • API Authentication Middleware — Bearer token authentication for all API routes. Set MC_API_TOKEN in .env.local to enable. Same-origin browser requests are automatically allowed.
  • Webhook HMAC-SHA256 Validation — Agent completion webhooks now require a valid X-Webhook-Signature header. Set WEBHOOK_SECRET in .env.local to enable.
  • Path Traversal Protection — File download endpoint now uses realpathSync to resolve symlinks and validate all paths are within the allowed directory.
  • Error Message Sanitization — API error responses no longer leak internal details (stack traces, file paths) in production.
  • Security Headers — Added X-Frame-Options, X-Content-Type-Options, Referrer-Policy, and Permissions-Policy headers via Next.js config.
  • Input Validation (Zod) — Request payloads for tasks, agents, and workspaces are validated with Zod schemas before processing.
  • Repository Audit — Purged sensitive files from git history, updated .gitignore to block database files and backups.

Added

  • Ed25519 Device Identity — Gateway pairing now uses Ed25519 key-based device identity for secure handshakes.
  • ARIA Hook — Real-time agent tracking bridge between ARIA and Mission Control (scripts/aria-mc-hook.sh).
  • Planning Poll Endpoint — New POST /api/tasks/[id]/planning/poll for long-poll planning updates.
  • Retry Dispatch — New POST /api/tasks/[id]/planning/retry-dispatch to retry failed task dispatches.
  • Auto-Dispatch Modulesrc/lib/auto-dispatch.ts for automatic task assignment after planning.
  • Planning Utilitiessrc/lib/planning-utils.ts with shared planning logic.
  • MC Bridge Scripts — Python and shell bridge scripts for external integrations.

Changed

  • Node.js v25 Support — Updated better-sqlite3 to v12.6.2 for Node v25 compatibility.
  • Default Port — Mission Control now defaults to port 4000 (previously 3000).
  • Improved Planning Tab — Enhanced UI with better question rendering, progress tracking, and error handling.
  • Agent Sidebar Improvements — Better status display, model selection, and agent management.
  • Activity Log Overhaul — Cleaner timeline UI with better type icons and formatting.
  • Live Feed Improvements — Better real-time event display with filtering options.

Fixed

  • Same-origin browser requests — Auth middleware no longer blocks the UI's own API calls.

1.0.1 - 2026-02-04

Changed

  • Clickable Deliverables - URL deliverables now have clickable titles and paths that open in new tabs
  • Improved visual feedback on deliverable links (hover states, external link icons)

1.0.0 - 2026-02-04

🎉 First Official Release

This is the first stable, tested, and working release of Mission Control.

Added

  • Task Management

    • Create, edit, and delete tasks
    • Drag-and-drop Kanban board with 7 status columns
    • Task priority levels (low, normal, high, urgent)
    • Due date support
  • AI Planning Mode

    • Interactive Q&A planning flow with AI
    • Multiple choice questions with "Other" option for custom answers
    • Automatic spec generation from planning answers
    • Planning session persistence (resume interrupted planning)
  • Agent System

    • Automatic agent creation based on task requirements
    • Agent avatars with emoji support
    • Agent status tracking (standby, working, idle)
    • Custom SOUL.md personality for each agent
  • Task Dispatch

    • Automatic dispatch after planning completes
    • Task instructions sent to agent with full context
    • Project directory creation for deliverables
    • Activity logging and deliverable tracking
  • OpenClaw Integration

    • WebSocket connection to OpenClaw Gateway
    • Session management for planning and agent sessions
    • Chat history synchronization
    • Multi-machine support (local and remote gateways)
  • Dashboard UI

    • Clean, dark-themed interface
    • Real-time task updates
    • Event feed showing system activity
    • Agent status panel
    • Responsive design
  • API Endpoints

    • Full REST API for tasks, agents, and events
    • File upload endpoint for deliverables
    • OpenClaw proxy endpoints for session management
    • Activity and deliverable tracking endpoints

Technical Details

  • Built with Next.js 14 (App Router)
  • SQLite database with automatic migrations
  • Tailwind CSS for styling
  • TypeScript throughout
  • WebSocket client for OpenClaw communication

0.1.0 - 2026-02-03

Added

  • Initial project setup
  • Basic task CRUD
  • Kanban board prototype
  • OpenClaw connection proof of concept

Roadmap

  • Multiple workspaces
  • Webhook integrations
  • API authentication & security hardening
  • Team collaboration
  • Task dependencies
  • Agent performance metrics
  • Mobile-responsive improvements
  • Dark/light theme toggle