Switchyard is a model-routing library and proxy that chooses which language model should handle each LLM request, based on factors such as cost and task suitability. It is for developers running LLM applications who want to route calls through existing gateways, custom harnesses, or a standalone proxy while retaining OpenAI and Anthropic client compatibility. The catalogue add-ons are workflows for operating or integrating Switchyard with coding agents and gateways.
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 skills/nvidia-nemo/switchyard/gold-reviewnpx skills add NVIDIA-NeMo/Switchyard --skill gold-reviewgit clone --depth 1 https://github.com/NVIDIA-NeMo/SwitchyardWrote 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/nvidia-nemo/switchyard/gold-review)<a href="https://agentmods.dev/skills/nvidia-nemo/switchyard/gold-review"><img src="https://agentmods.dev/badge/skills/nvidia-nemo/switchyard/gold-review.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 5.1 | $0.00074 | $0.02029 |
| Opus 5 | $0.00037 | $0.01014 |
| Sonnet 5 | $0.00015 | $0.00406 |
| Haiku 4.5 | $0.00007 | $0.00203 |
Grade A, and why
gold-review 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 6d 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 — 171 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CRAFT Search Gold Data Review
Systematic evaluation of gold answer quality for CRAFT Search benchmark tasks.
Trigger
User mentions a task ID like craft-httpx-9e2db24f or asks to review/investigate a specific task's gold data.
Critical Principle
Do NOT trust agent consensus. The agents are the subjects being measured — using their majority vote to validate the gold answer is circular reasoning. Agent output is a secondary signal at best. The primary evidence is YOUR OWN deep reading of the source code. If 7/7 agents report a function but the code shows it's irrelevant to the question, it stays out. If 0/7 agents report a function but the code shows it's the core mechanism, it goes in.
Workflow
1. Load task data
Run the loader script, which prints instruction, gold answer, tiers, and consensus counts:
uv run python .claude/skills/gold-review/scripts/load_task.py <task-id>
Read the output. Understand the question being asked — this anchors everything.
1.5. Trajectory Integrity Check
Run trajectory-based integrity checks against agent runs (if job dirs are available):
uv run python .claude/skills/gold-review/scripts/trajectory_integrity_check.py <task-id> \
--job-dirs <opus-job-dir> <codex-job-dir>
This checks for four issues (from the agentic-benchmark-eval checklists):
| Check | What it detects | Action if FAIL |
|---|---|---|
| Gold Contamination | Agent read /tests/gold_answer.json or /solution/solve.sh |
Task is invalid — recommend reject |
| Memorization | High reward (>0.5) with <3 tool calls | Gold may be too easy or leaked — investigate |
| Exploration Coverage | Agent claims files in answer.json that it never actually read | Agent may have hallucinated files — check gold scope |
| Answer Leakage | Instruction text contains gold file paths or private function names | Instruction needs adversarial rewrite |
If no job dirs are available, skip this step — the checks require agent trajectory data.
What ships with it
7 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.
- references/annotation-schema.md 2.1 KB
- references/investigation-rubrics.md 5.7 KB
- references/trajectory-integrity-checklist.md 3.0 KB
- scripts/apply_annotations.py 7.5 KB runs code
- scripts/load_task.py 4.8 KB runs code
- scripts/read_gold_functions.py 6.2 KB runs code
- scripts/trajectory_integrity_check.py 16 KB runs code
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.
- 6d ago First seen · 171 lines · 74 tokens per session scan A a58a34abf874
gold-review is a skill published in the GitHub repository NVIDIA-NeMo/Switchyard (2,713 stars, last pushed today), licensed Apache-2.0. It adds 74 tokens to every session and 2,029 once invoked, about $0.0004 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 skills, from other repositories
systematic-debugging
Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes.
local-ai-agents
Build local-first AI agents that run entirely on a developer workstation with Microsoft Foundry Local and Qwen function-calling models. Covers Small Language Models (SLMs), the OpenAI-compatible local endpoint, sandboxed local tools, local RAG with Chroma, local MCP servers, hybrid cloud/local routing, and the…
next-cache-components-adoption
Turn on Cache Components in a Next.js app and resolve the blocking routes it surfaces. Use when the user wants to enable, adopt, or migrate to Cache Components, flip the cacheComponents flag, work through a flood of blocking-prerender / instant validation errors, run the cache-components-instant-false codemod, or…
next-cache-components-optimizer
Drive a Next.js route to instant navigation by setting up an agentic loop, under Cache Components / PPR, on initial load (hard navigation) and client-side navigation (soft navigation). Encode the goal as a failing @next/playwright instant() e2e and work it to green, one verified route at a time; the shipped test then…
next-partial-prefetching-adoption
Turn on Partial Prefetching in a Next.js app and work through the insights it surfaces. Use when the user wants to enable or adopt Partial Prefetching, flip the partialPrefetching flag, opt routes in with export const prefetch = 'partial', audit Link prefetch={true} behavior, preserve existing prefetched UI with…
chronicle
Analyze Copilot session history for standup reports, usage tips, session search, and session reindexing. Use when the user asks for a standup, daily summary, usage tips, workflow recommendations, wants to search or find past sessions by keyword/file/PR, wants to reindex their session store, or asks about deleting…