Recipe Decompose

Orchestrator Definition

I am an orchestrator. I invoke a sub-agent for requirement analysis and decomposition, resolve the interaction from its evidence, and coordinate Linear registration. All Linear MCP operations are executed directly by the orchestrator. See references/register.md for the registration procedure.

Sub-agent: linear-prism:task-decomposer-linear

Input

$ARGUMENTS contains one of:

• A Linear issue URL (e.g., https://linear.app/team/issue/ENG-123)
• A file path to a PRD or requirement document
• Empty (ask the user which input to use)

Prerequisites

• Linear MCP connection is active. Verify it before starting; if unavailable, report the missing dependency and stop.

Workflow Overview

Round 1: Fetch input → Analyze outcome and scope → Continue or ask for a blocking user decision
Round 2: Propose decomposition strategy → Present task list → User adjusts
Round 3: Explicit approval of the current valid task list → Register in Linear

Detailed procedures for each round are in the references/ directory:

• Round 1: See references/analyze.md
• Round 2: See references/plan.md
• Round 3: See references/register.md

Scope Boundaries

In scope:

• Fetching and analyzing requirement content from Linear or local files.
• Resolving repository-local ambiguity from evidence and requesting only user-owned outcome or scope decisions.
• Decomposing requirements into typed tasks with dependency graphs.
• Registering tasks as Linear issues.

Out of scope:

• Modifying the source PRD or Linear issue directly.
• Generating Design Documents or Work Plans (downstream responsibility).
• Implementing any task produced by this skill.

Error Handling

| Condition | Action | |-----------|--------| | Linear MCP unavailable | Report the missing required dependency and stop | | Input URL returns empty or inaccessible content | Present the error, ask user for alternative input | | Agent returns recommended_action: needs_input | Ask only the returned user-owned questions, then re-analyze | | User rejects task list in Round 2 | Collect adjustment instructions, re-invoke agent in revise mode (see references/plan.md) |

Completion Criteria

• [ ] Every current in-scope desired behavior from the input maps to at least one task.
• [ ] Every shared prerequisite is required before the earliest dependent value unit can complete and verify its outcome.
• [ ] Each task has explicit service responsibility scopes.
• [ ] Dependency edges represent real execution or verification blockers.
• [ ] The exact task list approved by the user is registered in Linear.

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: shinpr
• Source: shinpr/linear-prism
• License: MIT

Install and usage instructions live in the source repository linked above.