agentleFS
sunholo-datasunholo-data/ailang/.claude/skills/collaboration-hub/SKILL.md

collaboration-hub

Skill38 starsChanged 9 days agoApache-2.0View on GitHub

Develop and modify the AILANG Collaboration Hub UI. Use when user asks to add features to the monitoring dashboard, modify the approval queue, update the message center, or make changes to the React frontend.

The file

We split this skill into the rules it gives an agent

9 rules. About 0 are common, 6 are distinctive, and 3 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈6

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 3

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Use when user asks to:

    line 33
  2. Distinctive:

    Add features to Control Plane dashboard

    line 34
  3. Distinctive:

    Update message center or approval queue

    line 36
  4. Distinctive:

    Add components to trace waterfall or event queue

    line 37
  5. Distinctive:

    Update button text to indicate handoff will occur

    line 122
  6. Very specific:

    Check hierarchy API: curl localhost:1957/api/observatory/tasks/TASK_ID/hierarchy

    line 183
  7. Very specific:

    Check spans exist: scripts/diagnose.sh spans TASK_ID

    line 184
  8. Distinctive:

    Use hierarchy API, not direct spans API:

    line 188
  9. Very specific:

    Update scripts/resources/SKILL.md

    line 223

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Collaboration Hub Developer

Build and modify the AILANG Collaboration Hub - a React-based UI for multi-agent coordination, task execution monitoring, and observability.

Quick Start

Starting services:

make services-start         # Start server + coordinator
make services-status        # Check both services
make services-stop          # Stop both
make services-restart       # Rebuild and restart

After UI changes:

make ui-deploy              # Build, clean, copy to server
make services-restart       # Restart with new build

Key URLs:

When to Use This Skill

Use when user asks to:

  • Add features to Control Plane dashboard
  • Modify Observatory or task hierarchy views
  • Update message center or approval queue
  • Add components to trace waterfall or event queue
  • Style or theme UI components
  • Diagnose hierarchy or span issues

Architecture Overview

Backend (Go):

  • internal/server/server.go - HTTP server, static files
  • internal/server/handlers_controlplane.go - Control plane API
  • internal/observatory/ - Telemetry and span storage
  • internal/observatory/hierarchy.go - Hierarchy building

Frontend (React + TypeScript + Vite):

ui/src/
├── App.tsx                     # Main app, navigation
├── features/
│   ├── controlplane/           # Control Plane v4 (main view)
│   │   ├── ControlPlane.tsx    # Main container
│   │   ├── components/         # MessageQueue, TraceWaterfall, etc.
│   │   └── hooks/              # useEventQueue, useTraceData
│   ├── observatory/            # Task and span exploration
│   ├── tasks/                  # Task execution monitoring
│   ├── agents/                 # Agent management
│   ├── messaging/              # Message center
│   └── approvals/              # Approval workflow
└── components/                 # Shared components

Database: SQLite at ~/.ailang/state/observatory.db

Available Scripts

scripts/diagnose.sh <command>

Diagnose Collaboration Hub and Observatory issues.

Commands:

.claude/skills/collaboration-hub/scripts/diagnose.sh health          # Server + DB check
.claude/skills/collaboration-hub/scripts/diagnose.sh task <id>       # Task details
.claude/skills/collaboration-hub/scripts/diagnose.sh spans <task_id> # Span hierarchy
.claude/skills/collaboration-hub/scripts/diagnose.sh orphans         # Find orphan spans
.claude/skills/collaboration-hub/scripts/diagnose.sh costs           # Cost breakdown
.claude/skills/collaboration-hub/scripts/diagnose.sh recent          # Recent activity

Key API Endpoints

Control Plane:

  • GET /api/controlplane/stats - Unified stats
  • GET /api/controlplane/exec-hierarchy - Execution hierarchy

Observatory:

  • GET /api/observatory/tasks/{id}/hierarchy - Full task hierarchy (RECOMMENDED)
  • GET /api/observatory/spans?task_id=X - Spans by task
  • GET /api/observatory/traces/{id} - Trace with all spans

Approvals (v0.6.5+):

  • GET /api/coordinator/pending - List pending approvals
  • POST /api/coordinator/approve/{id} - Approve (merge + optional handoff)
  • POST /api/coordinator/reject/{id} - Reject with feedback

For complete API reference: See resources/rest_api_reference.md

Approval Workflow (v0.6.5+)

The approval system uses unified approvals where merge and handoff are combined:

Approval TypeUI DisplayOn Approve
merge"Approve" buttonMerges code to dev
merge_handoff"Approve & Handoff" buttonMerges code AND triggers next agent

API response includes:

{
  "id": "apr-123",
  "type": "merge_handoff",
  "context_json": "{\"handoff_targets\":[\"sprint-planner\"],\"session_id\":\"...\"}"
}

UI considerations:

  • Show handoff targets when type === "merge_handoff"
  • Update button text to indicate handoff will occur
  • Show session continuity info (agent will resume with context)

Hierarchy & Trace Correlation

The hierarchy API applies three levels of virtual re-parenting:

  1. Timestamp Correlation - Fixes ailang.* spans under exec.tool_use
  2. Cross-Trace Merging - Links spans across different trace IDs
  3. Session-Based Merging - Links orphan traces by session.id

Why use hierarchy API? Child spans don't have task_id set. The hierarchy endpoint expands traces to include ALL spans.

For detailed algorithm: See resources/hierarchy_algorithm.md

Development Workflow

Adding a Component

// ui/src/features/controlplane/components/NewComponent.tsx
import styles from '../ControlPlane.module.css';

export const NewComponent: React.FC<Props> = ({ data }) => {
  return <div className={styles.newComponent}>{/* content */}</div>;
};

Adding Styles

/* Add to ControlPlane.module.css */
.newComponent {
  background: var(--bg-surface);
  padding: var(--space-4);
}

CSS Variables

/* Colors */
--bg-deep: #0a0c10;
--bg-surface: #161b22;
--primary: #25c2a0;
--success: #10b981;
--danger: #ef4444;

/* Text */
--text-primary: #e6edf3;
--text-secondary: #8b949e;

/* Spacing */
--space-1: 4px; --space-2: 8px; --space-3: 12px; --space-4: 16px;

Troubleshooting

Event shows no spans

  1. Run: scripts/diagnose.sh task TASK_ID
  2. Check hierarchy API: curl localhost:1957/api/observatory/tasks/TASK_ID/hierarchy
  3. Check spans exist: scripts/diagnose.sh spans TASK_ID

Missing child spans (exec.turn, etc.)

Use hierarchy API, not direct spans API:

# Wrong - only spans with task_id
curl "/api/observatory/spans?task_id=X"

# Correct - all spans in linked traces
curl "/api/observatory/tasks/X/hierarchy"

UI changes not appearing

make ui-deploy              # Clean build
# Hard refresh: Cmd+Shift+R

For database diagnostics: See resources/database_diagnosis.md

Resources

  • REST API Reference: resources/rest_api_reference.md
  • Database Diagnosis: resources/database_diagnosis.md
  • Hierarchy Algorithm: resources/hierarchy_algorithm.md
  • ID Relationships: resources/id_relationships.md - How IDs link across all 3 databases

Progressive Disclosure

  1. Always loaded: This SKILL.md (~200 lines)
  2. Execute as needed: Scripts in scripts/
  3. Load on demand: Resources in resources/

Self-Improvement

This skill self-improves. When issues are discovered:

  1. Identify the issue
  2. Update scripts/resources/SKILL.md
  3. Validate changes
  4. Log in CHANGELOG.md

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.