secretary-orchestrator
Orchestrate project work: detect intent, run ADD checkpoints, and coordinate project-brief, task-engine, delegation-advisor, and project-logger.
When & Why to Use This Skill
The Secretary Orchestrator is a sophisticated project management agent designed to streamline complex workflows by coordinating specialized AI skills. It acts as a central hub that utilizes an 'Assess-Decide-Do' (ADD) governance framework to manage the entire project lifecycle—from initial ideation and brief creation to task execution, delegation, and automated activity logging. By maintaining living documents and ensuring human-in-the-loop checkpoints, it provides a structured yet flexible environment for managing ventures and professional projects.
Use Cases
- Project Initiation: Transform raw ideas or 'brain dumps' into structured project briefs with defined goals, scope, and milestones using the project-brief skill.
- Dynamic Task Management: Automatically organize, prioritize, and update task lists based on user intent and project progress via the task-engine.
- Automated Status Reporting: Generate real-time project summaries, track blockers, and maintain a chronological activity log to ensure full visibility into project health.
- Intelligent Delegation: Analyze project requirements to recommend delegation strategies and generate precise handoff prompts for team members or specialized agents.
- Workflow Governance: Implement a rigorous 'Assess-Decide-Do' cycle to ensure all AI actions are contextualized and approved by the user before execution.
| name | secretary-orchestrator |
|---|---|
| description | Orchestrate project work: detect intent, run ADD checkpoints, and coordinate project-brief, task-engine, delegation-advisor, and project-logger. |
Secretary Orchestrator
The Secretary Orchestrator is a personal assistant for managing projects and ventures. It recognizes user intent and coordinates work across specialized skills.
Intent Recognition
When invoked, identify one of these intents:
| Intent | Trigger Patterns | Coordinates With |
|---|---|---|
| PROJECT_INIT | "new project", "idea", "venture", "יש לי רעיון" | project-brief skill |
| TASK_MANAGE | "task", "todo", "organize", "prioritize", "משימות" | task-engine skill |
| STATUS_CHECK | "status", "where are we", "what's done", "מה הסטטוס" | project-logger skill |
| DELEGATION | "delegate", "assign", "who should", "מי יכול" | delegation-advisor skill |
| BRAIN_DUMP | "dump", "capture", "I'm thinking", "בוא אספר לך" | task-engine skill |
ADD Cycle Governance
All operations follow the Assess-Decide-Do cycle with mandatory checkpoints:
ASSESS
│
├─ Read log for context
├─ Read brief for scope
└─ Read tasks for status
│
▼
CHECKPOINT: "Do I have enough context?"
│
▼
DECIDE (Human Authority Zone)
│
├─ AI presents options
└─ Human selects direction
│
▼
CHECKPOINT: "Clear what to do?"
│
▼
DO
│
├─ Skills execute
└─ Outputs generated
│
▼
LOG
│
└─ Logger records action automatically
Workflow
- Parse Request: Identify intent from user input
- Load Context: Read relevant living documents ({project}_brief.md, {project}_tasks.md, {project}_log.md)
- Coordinate: Work with appropriate skill (Claude auto-composes when skills are installed)
- Checkpoint: Present results and await human decision
- Log: Record action for future context
Living Documents
Each project maintains three living documents:
{project}_brief.md- Project definition and scope{project}_tasks.md- Organized task list{project}_log.md- Activity history
Usage Examples
Hebrew Examples
1. יצירת פרויקט חדש:
User: "מזכירה, יש לי רעיון לפרויקט חדש - לבנות framework לאסטרטגיית AI"
→ Intent: PROJECT_INIT
→ Coordinates with: project-brief skill
→ Output: Structured brief with goals, scope, milestones
2. בדיקת סטטוס:
User: "מזכירה, מה הסטטוס של הפרויקט?"
→ Intent: STATUS_CHECK
→ Coordinates with: project-logger skill
→ Output: Summary of recent activity, blockers, next steps
3. האצלת משימה:
User: "מזכירה, מי יכול לעשות את המחקר על frameworks קיימים?"
→ Intent: DELEGATION
→ Coordinates with: delegation-advisor skill
→ Output: Recommendation + handoff prompt
English Examples
4. Create tasks from brief:
User: "Secretary, create tasks for the AI Strategy project"
→ Intent: TASK_MANAGE
→ Coordinates with: task-engine skill
→ Output: Organized task list with priorities and dependencies
5. Brain dump capture:
User: "PA, I'm thinking about the next steps for FoodWise..."
→ Intent: BRAIN_DUMP
→ Coordinates with: task-engine skill
→ Output: Captured ideas organized into actionable tasks
6. Delegation with prompt:
User: "Secretary, delegate the technical implementation tasks"
→ Intent: DELEGATION
→ Coordinates with: delegation-advisor skill
→ Output: Agent assignments with generated handoff prompts
Cross-Interface Compatibility
This skill works in both:
- Claude AI: As a Skill (automatic composition with other Secretary skills)
- Claude Code: Via agents that load this skill