google / elephant-goldfish-model
git clone https://github.com/google/elephant-goldfish-model.git
An open-source method and skill set created at Google for spec-driven development with LLMs
+101
stars, 24h · #371
+~707
stars, 7 days · #309
+~3.0k
stars, 30 days · #281
193
★ total · +707% this week
16
forks · 17 open issues
Star history (as recorded here)
LanguagePython
LicenseApache-2.0
Created2026-09-13 (1mo ago)
Last pushtoday
Watchers4
OwnerOrganization
GitHub topics
agent-skillsai-agentsantigravityclaude-codeclaude-code-pluginclaude-skillscodexcodex-plugincoding-agentscontext-engineeringdeveloper-toolsgeminillmmulti-agentprompt-engineeringspec-driven-development
Links
- GitHub: github.com/google/elephant-goldfish-model
- Homepage: https://egm.dev
- Releases · Issues · Hacker News mentions
README excerpt
# Elephant-Goldfish Model (EGM) Elephant-Goldfish Model (EGM) is an open-source plugin for Spec-Driven Development with AI coding agents. Deep-context Elephant sessions draft specifications, and isolated zero-context Goldfish subagents stress-test them before you write code. ## Install | Harness | Install | | :--- | :--- | | **Any AI agent** | Ask your agent: `Install https://github.com/google/elephant-goldfish-model` | | **Antigravity** | `agy plugin install https://github.com/google/elephant-goldfish-model` (or `/plugin` in IDE) | | **Claude Code** | `claude plugin marketplace add google/elephant-goldfish-model && claude plugin install egm@egm` | | **Codex** | `codex plugin marketplace add google/elephant-goldfish-model && codex plugin add egm@egm` | | **Claude Cowork / ChatGPT** | See Harnesses for desktop and workspace setup | ## Quickstart 1. **Ideate & Stress-Test**: Run `/egm-brainstorm` to explore problem boundaries and test early concepts with a Goldfish swarm. 2. **Draft**: Run `/egm-draft` to convert the concept brief into a living design document under `docs/designs/ .md`. 3. **Review & Resolve**: Run `/egm-review docs/designs/ .md` to run the multi-stage Goldfish reviewer swarm. 4. **Submit & Implement**: Submit the verified design doc via pull request and implement code against the approved specification. In Claude Code, commands use the `/egm:` prefix (for example `/egm:egm-brainstorm`). For detailed setup guides and walkthroughs, see Getting Started, Harnesses, and Framework Architecture. ## Skills | Skill | Command | Description | | :---- | :------ | :---------- | | **Brainstorm** | `/egm-brainstorm` | Zero-doc ideation tool using a multi-lens Goldfish shootdown swarm to stress-test early concepts. | | **Draft** | `/egm-draft` | Iterative, section-by-section markdown design specification compiler. | | **Review** | `/egm-review` | Multi-stage verification pipeline orchestrator (Zoom Out -> Comprehension -> Critic -> Readiness). | | **Review Loop** | `/egm-review-loop` | Concurrent review subagent dispatch engine supporting direct and hierarchical dispatch. | | **Review Resolve** | `/egm-resolve` | Interactive triage and automated patch application engine for review findings. | | **Drift Report** | `/egm-drift-report` | Sequential AST scanner detecting divergence between checked-in specs and live code. | | **Bootstrap** | `/egm-bootstrap` | Recursive bottom-up codebase documentation generator and sentinel-block upserting engine. | ## Philosophy EGM uses two animal metaphors to structure AI agents across the software development lifecycle: * 🐘 **The Elephant (Remembers Everything)**: Deep-context agents load system history, architecture diagrams, component boundaries, and user journeys. The Elephant runs structured brainstorming, probes operational limits, and drafts comprehensive specifications. * 🐠 **The Goldfish (Zero-Context Memory)**: Stateless, isolated subagents read only the draft spec and live source code, without conversat