Works with the AI Agents you already use
devflow epic
Turns a PRD, TDD, designs or a prompt into an epic plus ordered tickets, each with a branch name and a stub plan. A critic checks coverage, dependencies and breaking changes before you review it.
$10
devflow epic
Example session with this skill installed
Break down the new auth-service PRD in docs/rfcs/auth.md into tickets for a new epic called 'user-auth-migration'.
- Read your context and instructions
- Compiled the devflow epic
I have analyzed docs/rfcs/auth.md and researched the existing identity-provider package. A draft epic has been created and verified by the plan critic.
Epic: user-auth-migration
| Ticket | Branch | Type | Description |
|---|---|---|---|
| T-1 | feat/auth-schema | migration | Define new Postgres schema for user sessions |
| T-2 | feat/auth-api-contract | contract | Export session validation types in shared-types |
| T-3 | feat/auth-service-logic | logic | Implement JWT rotation in auth-service |
| T-4 | feat/auth-ui-integration | ui | Update login component to use new auth-service endpoints |
Analysis Summary
- Sources:
docs/rfcs/auth.md,packages/identity-provider/src/** - Risks: High risk on T-1; session migration must be backward compatible.
- Decisions: Using JWT instead of opaque tokens to reduce DB lookups.
Next steps
- Review the full epic details in the plan mode buffer.
- Approve the plan to generate the
epic.mdand 4 ticket stubs. - Start the first ticket by running
git switch -c feat/auth-schema.
Connects securely to your tools. The creator never sees your data.
About this skill
Manually breaking a massive feature into implementation tickets is prone to dependency deadlocks and missed edge cases. This skill automates the transition from high-level PRDs and designs into a validated, ordered execution plan.
What it does
- Source analysis reads PRDs, TDDs, design files, and folder contexts to understand requirements.
- Dependency mapping orders tickets so that API contracts and migrations land before dependent UI work.
- Automated critique uses a subagent to check the draft for coverage gaps, breaking changes, and technical risks.
- Branch management generates unique branch names for every ticket and verifies they do not collide with existing work.
- Ticket scaffolding creates stub plans for every task, ready to be expanded into full implementation steps later.
How it works
- Input gathering where you provide the epic's scope via a prompt or reference files like
PRD.md. - Contextual research to identify affected services, existing patterns, and shared contracts.
- Epic drafting following a strict format that includes requirements tracing and a ticket table.
- Validation pass where an independent critic agent audits the plan against a comprehensive checklist.
- Final persistence saving the approved epic and individual ticket stubs to
docs/plan/.
Frameworks & tools
This skill requires git and bash. It is designed for agents with plan modes and subagent capabilities like Claude Code, using local scripts to verify branch availability and find existing plans.
Why this beats prompting it yourself
Generic prompts often miss cross-package dependencies or create tickets that are too large to review. This skill enforces a "critic" loop using a separate subagent context, ensuring a cold-eyed review of your architecture before a single line of code is written.
Use cases
- Decompose a 10-page PRD into a sequence of atomic, shippable PRs.
- Plan a feature that spans multiple microservices or monorepo packages.
- Ensure breaking API changes are ticketed and merged before frontend consumers.
- Create a standardized planning history in your repo without polluting git history.
Known limitations
Does not automatically create git branches or commit files. Requires a manual approval step before any files are written to disk.
Requires git and bash. Does not create branches or commit files. Works best in agents supporting subagents and plan modes.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before any payment.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed today
Requires