Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
git clone --depth 1 https://github.com/Buckeyes22/subagent-model-routingWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/commands/buckeyes22/subagent-model-routing/dag-routing)<a href="https://agentmods.dev/commands/buckeyes22/subagent-model-routing/dag-routing"><img src="https://agentmods.dev/badge/commands/buckeyes22/subagent-model-routing/dag-routing/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/commands/buckeyes22/subagent-model-routing/dag-routing"><img src="https://agentmods.dev/badge/commands/buckeyes22/subagent-model-routing/dag-routing.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00035 | $0.00354 |
| Opus 5.5 | $0.00014 | $0.00142 |
| Sonnet 5.5 | $0.00007 | $0.00071 |
| Haiku 4.5 | $0.00003 | $0.00035 |
Grade A, and why
dag-routing scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2mo ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
What it actually says
First invoke the subagent-model-routing skill (load it via the Skill tool), then follow it to orchestrate the task below.
Mechanism is mandatory — clear this before writing anything. If the task has dependency edges it is a DAG, and you MUST execute it via the Workflow tool (Workflow({ scriptPath })). Do NOT run a DAG as direct Agent({subagent_type:'…-shim'}) calls, direct Bash ~/.claude/scripts/*-shim.sh invocations, or inline work — that is the transport misroute (the work may reach the external models, but the DAG the user asked for never existed; "launched as shells" / "still on Opus"). If the task has no dependency edges, it is not a DAG — use the Part B flat-dispatch half of subagent-model-routing directly and say so.
Then follow the skill's gates: every node carries a namespaced codex, kimi, grok, or opencode shim agentType and model: 'sonnet', authored only through the codex() / grok() / kimi() / glm() / minimax() helpers (no bare agent()); run the misroute audit before launch; split judgment edges into separate workflow segments; pilot before fan-out.
Task: $ARGUMENTS
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 2mo ago First seen · 13 lines · 35 tokens per session scan A b189901c4a9f
dag-routing is a command published in the GitHub repository Buckeyes22/subagent-model-routing (5 stars, last pushed 2mo ago), licensed MIT. It adds 35 tokens to every session and 354 once invoked, about $0.0001 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other commands, from other repositories
swarm
Decompose task into parallel subtasks and coordinate agents.
ralph
Run ralph supervisor loop with Codex as executor.
hive
Query and manage swarm tasks (cells).
handoff
Properly end a swarm session - release reservations, sync state, generate continuation prompt.
inbox
Check swarm mail inbox for messages from other agents.
status
Check swarm coordination status - workers, messages, reservations.