Suraj1235/open-dynamic-workflows / packages/kimi-adapter/skills/odw
odw
VerifiedDynamic multi-agent workflows for Kimi Code CLI. Use when the user says "ultracode", "workflow:", "/deep-research", or wants broad multi-file work with planning, parallel execution, verification, or crash-resume.
Open Dynamic Workflows for Kimi Code CLI
Invoke this as /flow:odw <task> for the guided flow, or /skill:odw <task> to load the playbook directly.
The ODW MCP tools should already be available from ~/.kimi-code/mcp.json. The bridge scripts referenced below live next to this skill in scripts/.
flowchart TD
A([BEGIN]) --> B[Check whether the local ODW daemon and MCP tools are available. Run node scripts/daemon-bridge.js --check if tool availability is unclear.]
B --> C{Is ODW available?}
C -->|Yes| D[Run the user's task through ODW. Prefer the MCP tool odw_run, or use odw_plan first if the user asked to review the plan or the task is expensive/risky.]
C -->|No| E[Explain the missing daemon or auth token exactly, then use Kimi's native agent workflow as a fallback.]
D --> F[Report workflow id, topology, agent count, cost/time estimate, status, and final synthesized result.]
E --> F
F --> G([END])
Daemon path
- Prefer MCP tools when available: call
odw_health, thenodw_runfor direct execution orodw_planfirst when the user wants to review the plan, the task is expensive, or mutation risk is high. - If MCP tools are unavailable but the daemon is reachable, use the local bridge:
node scripts/daemon-bridge.js plan "<task>"node scripts/daemon-bridge.js exec plan.jsonnode scripts/daemon-bridge.js status <workflowId>node scripts/daemon-bridge.js result <workflowId>
- Report the workflow id, topology, agent count, cost/time estimate, and final synthesized result. Do not redo the same work manually while ODW is running it.
Kimi-native fallback
If the daemon is unavailable, state that full ODW requires the daemon, then use Kimi Code CLI's native planning and editing flow:
- Decompose into discovery -> parallel work -> adversarial verification -> synthesis.
- Keep independent work items narrow and structured.
- Verify the result before summarizing.
- Ask before executing risky mutations.
Safety
- Never include secrets in prompts, plans, logs, or workflow artifacts.
- Respect user approval and project safety rules before mutation.
- If ODW is unavailable, name the missing piece exactly instead of pretending the full engine ran.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 470k | Updated | Claude CodeCodex | Verified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |