epic-read

User Input

$ARGUMENTS

The input must contain a GitHub Epic (number or URL).

Objective

Read and understand the Epic. This skill only reads. It does not plan, write code, or create branches.

1. Read the Epic

Read:

• Epic title
• Epic description
• Objective
• Scope
• Out of Scope
• Architectural Context
• Acceptance Criteria
• Risks
• Work Items
• Dependencies and Execution Order
• Delivery Strategy
• Definition of Done
• Stop Conditions
• Linked discussions

The Epic structure is defined by tech-lead. When a section is missing, read the content wherever it is and report the missing section.

2. Read the Child Issues

Read every child Issue. For each one, capture:

• Number and title
• Outcome and Scope
• Acceptance criteria
• Declared dependencies on other Issues
• Implementation Context
• Architecture Map
• Current state (open / closed)

3. Check Clarity

If the Epic requirements are fundamentally unclear or contradictory, return BLOCKED with the exact questions that need answers.

Do not guess the intent of an unclear requirement.

4. Dependent Epics

If the Delivery Strategy declares Depends on Epic #<n>, check that Epic's pull request.

If it is not merged, return BLOCKED saying which Epic must be delivered first.

5. Final Output

Return:

• Status: READY or BLOCKED
• Epic number, title and objective
• Scope and Out of Scope
• Epic acceptance criteria
• Architectural Context, Risks and Stop Conditions
• Issues list (number, title, acceptance criteria, dependencies, Architecture Map, state)
• Declared execution order, if any
• Open questions (when BLOCKED)

Return this output to the calling skill and continue its flow. Do not end the run here unless this skill was invoked directly by the user.