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 skills add doublebearoliver-cyber/mihomo-ai-failover --skill dbear-mihomo-ai-failovergit clone --depth 1 https://github.com/doublebearoliver-cyber/mihomo-ai-failoverWrote 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/skills/doublebearoliver-cyber/mihomo-ai-failover/dbear-mihomo-ai-failover)<a href="https://agentmods.dev/skills/doublebearoliver-cyber/mihomo-ai-failover/dbear-mihomo-ai-failover"><img src="https://agentmods.dev/badge/skills/doublebearoliver-cyber/mihomo-ai-failover/dbear-mihomo-ai-failover/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/skills/doublebearoliver-cyber/mihomo-ai-failover/dbear-mihomo-ai-failover"><img src="https://agentmods.dev/badge/skills/doublebearoliver-cyber/mihomo-ai-failover/dbear-mihomo-ai-failover.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 | $0.00103 | $0.02705 |
| Opus 5 | $0.00051 | $0.01352 |
| Sonnet 5 | $0.00021 | $0.00541 |
| Haiku 4.5 | $0.00010 | $0.00270 |
Grade A, and why
dbear-mihomo-ai-failover 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 12d 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.
How it starts
The opening of the file, as written. The whole thing — 263 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Mihomo AI Provider failover
Use the local mihomo-ai-failover MCP tools. The CLI and daemon are
authoritative for live state and operations; this skill is the canonical
machine-facing workflow and safety contract.
OpenAI is enabled by default. Every other public Provider template is disabled until local evidence has been reviewed and the user authorizes a private overlay and persistent profile update.
When to use this skill
Use it for:
- diagnosing ChatGPT page, login, API, streaming, or Codex network failures;
- diagnosing WorkBuddy (China), Kimi, MiniMax, or Mavis network paths;
- discovering sanitized Provider hostnames on the user's Mac and previewing a narrow, private adaptation;
- inspecting the active, warm, and cold failover pools;
- checking whether the monitor switched correctly or entered backoff;
- previewing, installing, starting, stopping, rolling back, or uninstalling the local monitor.
Do not use it for general VPN tuning, fastest-node selection, global proxy changes, unrelated GitHub/npm/Docker failures, arbitrary website monitoring, non-Mihomo clients, or a remote Mac that the agent cannot reach through a trusted local MCP client.
Check runtime access before tools
This Skill is an operating contract, not the failover runtime itself. Before
following a tool workflow, check whether the trusted local
mihomo-ai-failover stdio MCP server and its tools are actually available.
- If the MCP tools are available, start with the read-only workflow below.
- If the tools are unavailable, do not claim that this Mac was diagnosed and do not silently install anything. Report the missing local dependency and direct the user to the repository root README for the pinned runtime or plugin installation.
- A cloud-only model cannot reach the user's Mihomo controller. It may explain the workflow, but it must stop before claiming live state or making changes.
Authority and trust
Use evidence in this order:
- live MCP results for current state;
- live tool descriptions and annotations for exact action semantics;
- this skill for workflow, boundaries, and stop conditions;
- human-facing repository documentation for background.
What ships with it
3 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- 12d ago First seen · 263 lines · 103 tokens per session scan A 80528c01b82e
dbear-mihomo-ai-failover is a skill published in the GitHub repository doublebearoliver-cyber/mihomo-ai-failover (0 stars, last pushed 1mo ago), licensed MIT. It adds 103 tokens to every session and 2,705 once invoked, about $0.0005 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 skills, from other repositories
torchtitan
Pretrain LLMs at scale with PyTorch 4D parallelism.
modal
Serverless GPU cloud for ML jobs and model APIs.
cloudflare-temporary-deploy
Deploy a Worker live, no account, via wrangler --temporary.
publish-site
Versioned site deploys to GitHub/Cloudflare/Netlify Pages.
tensorrt-llm
High-throughput LLM inference on NVIDIA GPUs.
stripe-projects
Provision SaaS services + sync creds via Stripe Projects.