model-routing
Choose the approved Luna or Sol lane before dispatching work.
Skill by siso-os
Install
Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.
npx agentmarket install skill/siso-os-model-routing
Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.
Get the weekly diff
The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.
About
Reviews
Write a reviewNo reviews yet. Installed it? Reviews come from verified installers only. How reviews work
Permissions
What this skill can do and what it asks for, from its SKILL.md. Checked Oct 25, 2026.
- Tools
- No tools found
- Secrets
- None asked for
- Install scripts
- None found
Network hosts it talks to 1
codeload.github.com· found in install command
Files it touches
~/.claude/skills/model-routing
What's inside
- Skill name
model-routing- Description
- Choose the approved Luna or Sol lane before dispatching work.
- License
- MIT
- Instructions
- 23 lines in SKILL.md · view on GitHub
Files (1)
- SKILL.md
SKILL.md (first lines)
# Model routing
Route by task shape and verification cost. A model label is not evidence that a
request succeeded.
| Need | Approved lane |
|---|---|
| Any normal Claude sub-agent task (exploration, implementation, tests, or synthesis) | Claude Agent `subagent_type: gpt-luna` |
| Any normal Codex delegated task | Native Codex Luna |
| Independent adversarial review or ship gate | Deliberate Sol reviewer (`gpt-sol`) |
| Tiny local operation | Current agent inline |
Luna is the default work lane. The one normal Claude call is
`Agent({subagent_type:"gpt-luna", prompt:"..."})`; do not pass a model override because
the frontmatter owns the pin. Use `gpt-sol` only when an independent cold review is
actually required. Do not use `model: haiku`, raw provider names, Herdr panes, or
ad-hoc launchers for ordinary Claude delegation. If a lane is unavailable, report
BLOCKED rather than silently substituting another lane.
For code discovery, use Serena `query_project` first and require exact evidence
before accepting a worker claim.From siso-os/siso-agent-playbook-public, MIT.
More from siso-os/siso-agent-playbook-public
- 0.565score
fast-build-loop
siso-os
How a product owner (one Opus integrator) drives up to 12 Sol builders to finish a product fast - lanes, a 20-minute merge train, deploy the last green SHA…
SkillAFreshness: AClaude Code +9 - 0.561score
fan-out
siso-os
Plan a small parallel wave for genuinely independent units with explicit ownership, compact returns, and one convergence step.
SkillAFreshness: AClaude Code +9 - 0.561score
subagents
siso-os
Dispatch bounded sub-agents with compact evidence-backed returns — Luna roles in Codex, Claude sub-agents in Claude Code — and keep consequential judgment…
SkillAFreshness: AClaude Code +9 - 0.558score
haiku-dispatch
siso-os
Retired compatibility pointer for old Haiku/MiniMax dispatches; use the GPT-5.6 Luna Claude Agent instead.
SkillAFreshness: AClaude Code +9 - 0.556score
telemetry
siso-os
Inspect lane health and tokens-per-delivered-task before a parallel wave or after a failed dispatch; report evidence without choosing a replacement model.
SkillAFreshness: AClaude Code +9
Signals
- Stars
- 0
- Forks
- 0
- Open issues
- 0
- Last commit (this folder)
- Oct 20, 2026 (8 days ago)
From GitHub (fetched Oct 24, 2026).
Details
- License
- MIT
- First seen
- Oct 24, 2026
- Last checked
- Oct 24, 2026
- Source
- GitHubgithub-graphql
More from siso-os
- 0.565score
fast-build-loop
siso-os
How a product owner (one Opus integrator) drives up to 12 Sol builders to finish a product fast - lanes, a 20-minute merge train, deploy the last green SHA…
SkillAFreshness: AClaude Code +9 - 0.561score
fan-out
siso-os
Plan a small parallel wave for genuinely independent units with explicit ownership, compact returns, and one convergence step.
SkillAFreshness: AClaude Code +9 - 0.561score
subagents
siso-os
Dispatch bounded sub-agents with compact evidence-backed returns — Luna roles in Codex, Claude sub-agents in Claude Code — and keep consequential judgment…
SkillAFreshness: AClaude Code +9 - 0.558score
haiku-dispatch
siso-os
Retired compatibility pointer for old Haiku/MiniMax dispatches; use the GPT-5.6 Luna Claude Agent instead.
SkillAFreshness: AClaude Code +9 - 0.556score
telemetry
siso-os
Inspect lane health and tokens-per-delivered-task before a parallel wave or after a failed dispatch; report evidence without choosing a replacement model.
SkillAFreshness: AClaude Code +9 - 0.529score
analyzing-video
siso-os
Extract frames or scenes from a video with ffmpeg and analyse them in parallel.
SkillBFreshness: BClaude Code +9
Cite this page
model-routing skill. Agent Mart. https://agentmart.dev/skills/siso-os-model-routing (last updated 2026-10-24)