notion-ops
Notion Ops
Operate Notion as a reliable execution system for multi-agent and human workflows.
Core Outcomes
- Keep
Total tasksas the single source of execution truth. - Enforce immutable
Work Logsfor every status transition. - Enforce signature policy for AI and human actors.
- Preserve deterministic continuation across MCP clients.
- Keep handoff, context, and approval chains auditable.
Execution Workflow
1. Audit Baseline
- Verify required databases and relations.
- Verify role model (
Orchestrator,Planner,Executor,Reviewer,Human Supervisor). - Verify template quality and required task fields.
- Verify continuity chain (
Runs,Handoffs,Context Packets).
Use: references/database-contracts.md
2. Harden Template Layer
- Keep one canonical SOP template for governance.
- Maintain specialized templates (AI solo, incident, research, supervisor).
- Ensure template defaults are safe:
Status=BacklogSignature Status=Missingfor AI-capable templatesInteroperability=Ready- explicit
Policy Version
Use: references/operating-model.md
3. Operate Task Lifecycle
- Fill mandatory fields before
Ready. - Create first work log before
In progress. - Attach evidence before review gates.
- Enforce signatures and approvals before
Done. - If blocked, publish blocker owner and next action immediately.
Use: references/checklists.md
4. Maintain MCP Continuity
- On pause/transfer, publish one
Handoffsrecord. - Keep one
Context Packetsrecord markedCurrent. - Include executable
Resume CommandandNext Action. - Link active task to client profile and context/handoff records.
Use: references/mcp-continuation.md
5. Run Daily Control
- Check signature compliance.
- Check transition-log completeness.
- Check unresolved blockers and aging
In progresstasks. - Check open approvals and stale context packets.
Non-Negotiable Rules
- Never move AI-executed work to
Donewithout valid AI signature. - Never change task status without a matching
Work Logsentry. - Never transfer work across clients without handoff + context update.
- Never close incidents without mitigation proof and follow-up actions.
Output Contract
When reporting work, return:
Current State: concise health snapshot.Actions Performed: exact updates and affected artifacts.Compliance Gaps: signature/log/approval/continuity misses.Next Actions: ordered commands for immediate continuation.
Failure Handling
- Relation mismatch: fetch schema and map exact property names before writing.
- Ambiguous selection update: use unique snippets or replace full content.
- Rate limits: switch to direct page fetches by ID and continue incrementally.
- Partial continuity data: freeze status advancement, then repair handoff/context links.
References
More from alicoder001/agent-skills
reasoning
Chain-of-thought reasoning, self-reflection, and systematic problem-solving patterns for AI agents. Use before any complex task to ensure logical and accurate solutions.
38typescript
TypeScript strict mode patterns, naming conventions, and type safety rules. Use when writing TypeScript code, defining types, or reviewing TypeScript projects. Includes generics, utility types, and best practices.
35collaboration
Multi-agent communication, task delegation, and coordination patterns. Use when working with multiple agents or complex collaborative workflows.
27solid
SOLID, DRY, KISS, and clean code principles for TypeScript applications. Use when designing scalable architecture, writing maintainable code, or reviewing code quality.
25security
Security best practices for web applications. Use when handling user input, authentication, or sensitive data. Covers XSS, SQL injection, CSRF, environment variables, and secure coding patterns.
22memory
Working memory management, context prioritization, and knowledge retention patterns for AI agents. Use when you need to maintain relevant context and avoid information loss during long tasks.
22