ido4
Your AI agents write code. ido4 gives them shared understanding, institutional memory, and quality enforcement.
ido4 is the platform that makes AI-hybrid software development work at scale. It sits inside your AI coding environment and gives every session full project context — what to build, what was built before, who depends on your output. Task intelligence recommends the highest-leverage work. Deterministic governance validates every action. Institutional memory compounds knowledge across sessions.
> What should I work on?
get_next_task { agentId: "agent-alpha" } Recommended: #42 Auth token rotation (score: 72) Unblocks 3 downstream. Epic is 60% complete — finish it.
> Load context for #42
get_task_execution_data { issueNumber: 42 } ✓ Upstream: #38 built JWT refresh (RSA-256, 30min TTL) ✓ Siblings: #39, #40 established error handling pattern ✓ Downstream: #45, #47, #51 waiting on this task
> Start the task
start_task { issueNumber: 42 } ✓ BRE: 6/6 steps passed. Transitioning to In Progress.The agent didn’t start from scratch. It knew what to work on, why it matters, what was built upstream, and who depends on its output. That’s ido4.
Why ido4 exists
AI coding agents are powerful but stateless. Each session starts fresh — no memory of what other agents built, no understanding of your methodology, no awareness of dependencies or deadlines.
ido4 bridges this gap with four capabilities:
- Context intelligence — one call assembles task spec, upstream decisions, sibling patterns, downstream consumers, risk flags. No agent starts blind.
- Task intelligence — 4-dimension scoring (cascade value, epic momentum, capability match, freshness) recommends the highest-leverage work for each agent.
- Quality enforcement — 34 deterministic validation steps enforce your workflow. Dependencies, integrity rules, quality gates — real code, not LLM instructions.
- Institutional memory — structured context comments capture what was built, what was decided, what interfaces were created. The next agent inherits accumulated understanding.
Plus multi-agent coordination (task locking, work distribution, handoff), full audit trail, compliance scoring, and three built-in methodologies.
Pick your methodology
ido4 is methodology-agnostic. The engine reads profiles; profiles define everything.
Hydro — Wave-based delivery with epic integrity. All tasks in an epic ship together. Five governance principles. Built for AI-augmented consulting teams.
Scrum — Sprint-based execution with type-scoped pipelines. User stories need acceptance criteria. Bugs skip dependency checks. Spikes have relaxed DoD. Tech debt requires 2 reviewers.
Shape Up — Fixed time, variable scope. Six-week cycles with a circuit breaker — if a bet doesn’t ship by deadline, it’s killed. No extensions. Scope gets hammered, not timelines.
> Initialize ido4 with Scrum for my-org/my-projectThat’s it. ido4 creates the GitHub Project, sets up fields and statuses matching your methodology, and starts governing.
What you get
58 MCP tools (Hydro) / 56 (Scrum) / 54 (Shape Up) — generated dynamically from your methodology profile. Every write tool validates through the BRE, supports dry-run, and creates an audit entry.
21 skills — intelligent governance workflows. Morning standup that spots cascade blockers. Sprint planning that enforces Definition of Ready per work item type. Retrospectives with real cycle time data. Zero-friction onboarding with auto-clone demo, guided four-act demo walkthrough, and interactive sandbox exploration.
4 agents — a PM agent with persistent memory, a code analyzer for spec decomposition, a technical spec writer, and a spec reviewer.
Quick links
| I want to… | Go here |
|---|---|
| Get running in 5 minutes | Installation |
| See governance in action | Sandbox Demo |
| Understand the methodology options | Methodologies |
| See all available tools | Tool Reference |
| Learn how the BRE works | Business Rule Engine |
| Set up multiple AI agents | Multi-Agent Governance |
| Configure for enterprise | Enterprise Features |
| Look up a term | Glossary |