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 instructions/ypollak2/llm-router/copilot-instructionsgit clone --depth 1 https://github.com/ypollak2/llm-routerWrote 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/instructions/ypollak2/llm-router/copilot-instructions)<a href="https://agentmods.dev/instructions/ypollak2/llm-router/copilot-instructions"><img src="https://agentmods.dev/badge/instructions/ypollak2/llm-router/copilot-instructions.svg" alt="Measured on agentmods" 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 | $0.00850 | $0.00850 |
| Opus 5 | $0.00425 | $0.00425 |
| Sonnet 5 | $0.00170 | $0.00170 |
| Haiku 4.5 | $0.00085 | $0.00085 |
Grade A, and why
llm-router copilot-instructions.md 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 2d 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 — 92 lines — stays where its author put it; the contents beside it link to each section on GitHub.
LLM Router — VS Code / GitHub Copilot Routing Rules
These rules apply when using llm-router MCP tools inside VS Code with GitHub Copilot. VS Code loads MCP servers from .vscode/mcp.json (workspace) or the user mcp.json. Use
llm_autofor routing + cross-session savings tracking.
How to Route
VS Code loads MCP servers on startup. Once llm-router is registered, call tools by name:
Rule: for research/generate/code/analyze tasks → call llm_auto(prompt=<the task>)
Add to .github/copilot-instructions.md for workspace-level routing guidance:
When a task requires live web search, call the llm_research MCP tool.
For code generation or complex analysis, call llm_code or llm_analyze.
For auto-routing with savings tracking, call llm_auto.
Task Type → Tool Mapping
| Task | Tool | Notes |
|---|---|---|
| Simple factual question | llm_query |
Gemini Flash / Groq |
| Research / current events | llm_research |
Perplexity (web-grounded) |
| Writing / content | llm_generate |
Gemini Flash / Haiku |
| Deep analysis | llm_analyze |
GPT-4o / Gemini Pro |
| Code generation | llm_code |
Ollama → Codex → o3 |
| Unknown / auto-detect | llm_auto |
Classifies + routes |
Savings Visibility
Run llm_savings in Copilot Chat to see cross-session totals.
Token-Efficient Responses
Skip preamble. Lead with result. Fragments fine when meaning is clear. No trailing summaries. ≥3 items → bullets. Never restate the user's request.
LLM Router — VS Code / GitHub Copilot Routing Rules
LLM Router routes each task to the cheapest capable model before Copilot generates its own response. Call the appropriate tool BEFORE answering.
Pull routing: Copilot uses pull routing — YOU must call the tool. The tool returns the response; relay it verbatim. Requires VS Code ≥ 1.99 in agent mode.
Task Type → Tool Mapping
| Task | Tool | Notes |
|---|---|---|
| Simple factual question | llm(task="query") |
Ollama → Gemini Flash → GPT-4o-mini |
| Research / current events | llm(task="research") |
Perplexity (web-grounded) |
| Writing / content | llm(task="generate") |
Gemini Flash → Haiku |
| Deep analysis / review | llm(task="analyze") |
Ollama → GPT-4o |
| Code generation / refactor | llm(task="code") |
Ollama → Codex → o3 |
| Deep reasoning / proofs | llm(task="analyze") |
Extended-thinking model |
| Unknown / auto-detect | llm_auto |
Classifies + routes automatically |
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.
- 2d ago Changed a06392337fc0
- 5d ago First seen · 92 lines · 850 tokens per session scan A 151e3cbe067f
llm-router copilot-instructions.md is an instructions file published in the GitHub repository ypollak2/llm-router (77 stars, last pushed today), licensed MIT. It adds 850 tokens to every session, about $0.0042 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 2026-08-30.
Other instructions, from other repositories
ClawRouter CLAUDE.md
Claude Code instructions for BlockRunAI/ClawRouter, covering clawrouter, commands, project structure, key dependencies and conventions.
botmux CLAUDE.md
Claude Code instructions for deepcoldy/botmux, covering botmux, 构建 & 运行, bun 开发链路, worktree 的 nodemodules：共享还是独立（改前必读） and 编译态（单文件二进制）注意.
codedb AGENTS.md
AGENTS.md instructions for justrach/codedb, covering codedb agent guidelines, what codedb is (and isn't), review guidelines, pre-merge verification and security-sensitive areas.
happier CLAUDE.md
Claude Code instructions for happier-dev/happier, covering claude code and typescript commands.
InvestSkill GEMINI.md
Gemini CLI instructions for yennanliu/InvestSkill, covering investskill — gemini cli setup & usage guide, installation & setup, quick start, navigate to the investskill directory and start gemini cli (loads gemini.md automatically).
aisix CLAUDE.md
Claude Code instructions for api7/aisix, covering claude.md, 1. think before coding, 2. simplicity first, 3. surgical changes and 4. goal-driven execution.