Anneal Umbrella Dev
Turn a vague task into a scored, red-teamed plan artifact (XML prompt plus plan directory) before you commit to implementation.
Overview
anneal-umbrella-dev is a plugin marketplace for the Validate phase that ships three Claude Code plan-review variants producing XML prompts and plan directories via simulated annealing.
What is this marketplace?
- Three installable plugin variants (umbrella dev ships 3 plugins) for side-by-side architecture comparison
- Simulated annealing loop: heat candidates, cool scores, temper with red team until plan quality converges
- Named review archetypes Metis, Momus, and Oracle plus always-on red team critique
- Outputs XML prompt, plan directory, and skill enrichment verified with validate-xml.py
- 9 Greek-god cast pipeline verified end-to-end in Claude Code workers per project verification summary
- Marketplace bundles 3 plugins across installable architecture variants
- Verification summary references 9 Greek-god agents in the cast pipeline with end-to-end Claude Code worker proof
Community signal: 1 GitHub stars.
What problem does it solve?
Vague tasks become brittle one-shot plans because nobody systematically scores, prunes, and red-teams options before agents start coding.
Who is it for?
Solo builders using Claude Code who want comparable plan-review architectures and hard artifacts before implementation.
Skip if: Quick edits or single-file codegen jobs where formal XML plans and multi-agent Greek-god dispatch are overhead.
What do I get? / Deliverables
After install, you converge on a validated XML prompt and plan folder with critique traces suitable for downstream execution skills.
- Validated XML prompt suitable for downstream agent execution
- Plan directory with enrichment aligned to chosen Anneal variant
- Documented critique trail from red team and named review archetypes
Plugins in this marketplace
3 plugins — install individually after you add the marketplace.
Recommended Marketplaces
Journey fit
Canonical shelf is Validate because Anneal’s core job is tightening scope and plan quality before full build execution. Scope fits simulated-annealing plan review, Metis Momus Oracle archetypes, and convergence on an execution-ready artifact.
How it compares
Multi-variant planning methodology marketplace, not a lightweight lint or code-review skill.
Common Questions / FAQ
Who is Anneal Umbrella Dev for?
Advanced Claude Code users who treat planning as a first-class product step and want Metis, Momus, and Oracle style review with red teaming.
When should I use Anneal Umbrella Dev?
When a task is underspecified and you need scored candidates, temper cycles, and a validate-xml-ready artifact before build.
How do I add Anneal Umbrella Dev to my agent?
Install the Anneal umbrella marketplace in Claude Code, pick one of the three plugin variants to test, then invoke the documented cast or plan pipeline on your task and validate XML output locally.