# MCP Linear Planner

> agentic-workflows-blueprint.workflow.mcp-linear-planner

- Skill: `devton/mcp-linear-planner` (Agent Skill)
- Install (CLI): `npx skillmds@latest add devton/mcp-linear-planner`
- Raw SKILL.md: https://api.skillmd.com/api/skills/devton/mcp-linear-planner/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: AI & ML
- Author: devton (https://skillmd.com/u/devton)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/devton/mcp-linear-planner

---

## agentic-workflows-blueprint.workflow.mcp-linear-planner

### Goal

Validate MCP prerequisites and produce a deterministic execution plan for
Linear operations.

### Scope

- Applies to: preparing any Linear MCP integration run.
- Does not cover: executing create/update actions in Linear.

### Triggers

- "Plan Linear sync via MCP"
- "Prepare MCP workflow for Linear"
- "Validate Linear MCP readiness"

### Inputs

- `mcpServer` (default: `user-Linear`)
- `intent` (create project items, update statuses, organize milestones)
- `projectContext` (optional ids/names already known)
- `baseBranch` (optional, for linking implementation scope)
- `changeWindow` (optional): allowed execution window for external-state writes
- `riskClass` (optional): risk tag to drive stricter preflight behavior
- `rollbackTicket` (optional): external rollback/change reference

### Invariants

- Always inspect tool schema/descriptor before calling any MCP tool.
- If server exposes `mcp_auth`, run auth before operational calls when needed.
- Keep parameter names aligned with tool schema (no guessed fields).
- Produce a written action plan before mutating external state.
- If `riskClass` is `high`, include explicit hold point for operator approval before execution.

### Procedure

1. List/inspect available Linear MCP tools and required arguments.
2. Verify auth readiness; if blocked, route to auth and retry.
3. Resolve target team/project context (or mark as required input).
4. Build an operation plan:
  - ordered tool calls;
  - payload shape per call;
  - idempotency/safety notes;
  - operational metadata mapping (`changeWindow`, `riskClass`, `rollbackTicket`) when provided.
5. Emit a plan artifact to be consumed by `mcp-linear-sync`.

### Outputs

- `linearMcpPlan` with ordered steps and payload schemas.
- `preflightChecklist` (`tools`, `auth`, `context`) marked pass/fail.
- `blockingItems` list when execution cannot continue.

### Review gate

- Tool schema was explicitly checked before planned calls.
- Auth path is defined (or confirmed not required).
- Plan has deterministic call order and parameter mapping.
- High-risk plans include explicit human approval hold point.

### References

- `../../SKILL.md`
- `../mcp-linear-sync/SKILL.md`
- [Interactive HTML View](./README.html)