[Feature]: Add /omo-spec — A Unified Spec‑Driven Workflow for Oh‑My‑OpenCode #854
Description
Prerequisites
- I have searched existing issues and discussions to avoid duplicates
- This feature request is specific to oh-my-opencode (not OpenCode core)
- I have read the documentation
Problem Description
Oh‑My‑OpenCode (OMO) provides an extremely powerful multi‑agent execution environment, including:
- The Sisyphus orchestrator with multi‑model reasoning
- Specialist agents like oracle, librarian, and explore
- 20+ workflow hooks (todo‑continuation‑enforcer, session-recovery, context-window-monitor, etc.)
- Ralph‑Loop, enabling autonomous iteration until
<promise>DONE</promise>
However, OMO currently lacks a native specification lifecycle.
Tools like GitHub Spec‑Kit (Reference: https://github.com/github/spec-kit — Spec Kit documentation & CLI) do support OpenCode, but they:
- Are designed around template-driven workflows rather than multi-agent orchestration
- Require multi-step manual commands (
/speckit.specify,/speckit.clarify,/speckit.plan,/speckit.tasks) - Do not integrate with OMO’s agents, hooks, MCPs, or Ralph‑Loop
- Cannot leverage Sisyphus's planning capabilities, nor the oracle/librarian/explore agents
- Cannot automatically progress from specification → planning → implementation
This creates friction for OMO users who want:
- A guided “What & Why” discovery phase
- Structured specification artifacts
- Automatic technical planning
- A clean task breakdown
- A seamless handoff into Ralph‑Loop execution
OMO currently delivers world‑class execution, but users are left without structured upstream scaffolding.
Proposed Solution
Introduce a single unified slash command:
👉 /omo-spec "<project or feature description>"
This command would orchestrate the entire spec-driven lifecycle:
1. Specification Discovery
- Ask guided “What & Why” questions
- Produce a structured
.omo/spec/spec.md - Capture user stories, success criteria, constraints, edge cases
2. Clarification
- Identify ambiguities
- Ask up to three clarification questions
- Update the spec automatically
3. Technical Plan Generation
- Use Sisyphus + oracle + librarian
- Produce
.omo/spec/plan.md - Include architecture, modules, workflow, design decisions
4. Task Breakdown
- Convert the plan into
.omo/spec/tasks.md - Order tasks by dependency
- Identify parallelizable steps
5. Optional Execution
Prompt user to start implementation via Ralph‑Loop:
/ralph-loop "<plan + tasks>" --completion-promise="DONE"
This produces a seamless flow:
Idea → Spec → Clarification → Plan → Tasks → Implementation (Ralph‑Loop).
All with one command.
Alternatives Considered
Using Spec‑Kit directly
- Spec‑Kit does work with OpenCode, but it does not integrate with:
- Sisyphus multi‑agent orchestration
- OMO’s hooks
- Ralph‑Loop execution
- OMO’s tooling ecosystem
- It relies on multi-step user commands and git scripts
- It is template-driven, not agent-driven
- It cannot interact with OMO’s runtime or agent context
Manual prompting
Users can manually write specs, plans, and tasks, but this:
- Causes drift
- Requires prompt-engineering skill
- Produces inconsistent quality
- Does not leverage OMO’s agents effectively
- Yields poor results when fed into Ralph‑Loop
A native OMO solution is significantly more powerful and user-friendly.
Doctor Output (Optional)
Additional Context
This feature would unify OMO’s strengths:
- Multi-agent reasoning
- Planning
- Code generation
- Analysis
- Autonomous execution (Ralph‑Loop)
While also providing a structured developer experience similar to Spec‑Kit — adapted to OMO’s architecture.
A native /omo-spec command would dramatically improve developer workflows and reduce friction, especially for users working on complex multi-file or multi-phase projects.
References
- GitHub Spec‑Kit: https://github.com/github/spec-kit
- DeepWiki documentation: https://deepwiki.com/github/spec-kit
Feature Type
New Tool
Contribution
- I'm willing to submit a PR for this feature
- I can help with testing
- I can help with documentation