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.
npx agentmods add commands/pealmeida/anymodel-plugin/delegate-withgit clone --depth 1 https://github.com/pealmeida/anymodel-pluginWhat 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 | $0.00009 | $0.00527 |
| Opus 5 | $0.00005 | $0.00264 |
| Sonnet 5 | $0.00002 | $0.00105 |
| Haiku 4.5 | $0.00001 | $0.00053 |
Grade A, and why
delegate-with 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 yesterday.
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
Delegate a task after helping the user choose the provider/model. This is the delegation-focused semantic chain for users who know they want execution, but not necessarily which model to use.
Raw user request:
$ARGUMENTS
Selection rules:
- If the request contains
--model <provider/model>, preserve it and skip model selection. - If it contains a natural model token such as
zai/glm-5.2,ollama/qwen3-coder, oropencode-go/kimi, convert that to--model <token>. - If it contains only a provider id, run:
Then ask the user to choose one reachable model returned for that provider.node "${CLAUDE_PLUGIN_ROOT}/scripts/companion.mjs" models --provider <provider> --json - If it contains no provider/model, run:
Then ask the user to choose from up to three reachablenode "${CLAUDE_PLUGIN_ROOT}/scripts/companion.mjs" models --json<provider>/<model>options. - Ignore providers where
okis false or no models are returned. - If no reachable model exists, return the models output verbatim and tell the user to run
/anymodel:setup.
Task rules:
- Preserve runtime flags exactly:
--engine,--bridge,--write,--background,--wait,--resume,--fresh,--cwd. - If no task remains after removing provider/model selector words, ask what the selected model should investigate, solve, or implement.
- Do not rewrite the user's task.
Dispatch:
- Invoke the
anymodel:anymodel-runnersubagent via theAgenttool (subagent_type: "anymodel:anymodel-runner"), forwarding the final request as if the user had called/anymodel:delegate. - The final request must include
--model <selected provider/model>. - The subagent must make exactly one Bash call to:
node "${CLAUDE_PLUGIN_ROOT}/scripts/companion.mjs" delegate <final arguments> - Return the companion stdout verbatim. Do not summarize or add commentary.
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.
- yesterday First seen · 45 lines · 9 tokens per session scan A 031710dfc546
delegate-with is a command published in the GitHub repository pealmeida/anymodel-plugin (0 stars, last pushed 24d ago), licensed Apache-2.0. It adds 9 tokens to every session and 527 once invoked, about $0.0000 per session on Opus 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
design
Design-space exploration tree. Wraps /cc-tree:tree with the design preset preselected — enumerates options × trade-offs × reversibility × cost × fit-with-constraints for an engineering / product / process decision. Produces a comparison table and a RECOMMENDED short-list. 中文触发：设计空间探索、方案对比与取舍、给出推荐短名单。.
step-research
Always research before proposing a fix. The Untether bug you're chasing is often a known upstream engine quirk, a previously-fixed regression, or a documented config gotcha.
step-classify
This is the canonical Untether issue-class table. Pick exactly one primary class per issue (the area where the root cause lives). If behaviour crosses boundaries, note a secondary class.
README
This directory is the canonical source for each step of Untether's 8-step deep debugging protocol. The parent command file at ../debug.md orchestrates the steps; each file here owns the detail for one step (or one shared output).
axstream-teach
Compile a UI task into a fast replayable macro (do it once, instant forever).
auto
AI auto strategy executor. Analyzes task and automatically selects and executes the best strategy: direct execution, hybrid-auto PRD generation, hybrid-worktree isolated development, or mega-plan multi-feature orchestration.