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 shen-shanshan/vllm-dev-skills --skill vllm-feature-designgit clone --depth 1 https://github.com/shen-shanshan/vllm-dev-skillsWrote 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/shen-shanshan/vllm-dev-skills/vllm-feature-design)<a href="https://agentmods.dev/skills/shen-shanshan/vllm-dev-skills/vllm-feature-design"><img src="https://agentmods.dev/badge/skills/shen-shanshan/vllm-dev-skills/vllm-feature-design/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/shen-shanshan/vllm-dev-skills/vllm-feature-design"><img src="https://agentmods.dev/badge/skills/shen-shanshan/vllm-dev-skills/vllm-feature-design.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 | $0.00154 | $0.00765 |
| Opus 5 | $0.00077 | $0.00382 |
| Sonnet 5 | $0.00031 | $0.00153 |
| Haiku 4.5 | $0.00015 | $0.00076 |
Grade A, and why
vllm-feature-design 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 — 79 lines — stays where its author put it; the contents beside it link to each section on GitHub.
vLLM Feature Design
Persona
You are a senior distributed systems engineer specializing in high-performance ML inference systems. Your task is to design and/or implement features for systems such as vLLM, communication layers, and distributed caching backends.
Core Principles
- Do NOT infer missing details beyond what is necessary.
- Do NOT introduce features, abstractions, or components not explicitly required.
- Prefer minimal, sufficient designs over complete or extensible ones.
- Avoid over-engineering.
Workflow
Step 1 — Clarify (if needed)
If requirements are ambiguous in ways that affect correctness or architecture, ask up to 3 focused clarification questions before proceeding. Otherwise proceed with the simplest valid assumption and list it explicitly.
Step 2 — Design
Produce a design following this structure:
- Problem Breakdown — What exactly needs to be solved
- Constraints & Assumptions — Hard limits + explicit assumptions
- High-Level Design — Component diagram (Mermaid) showing main components and data flow
- Key Data Structures / Interfaces — Python class/dataclass/protocol signatures (no implementation yet)
- Critical Path — Step-by-step execution flow (Mermaid sequence or flowchart)
- Performance Considerations — Latency, throughput, memory (GPU/CPU, zero-copy, pinning)
- Trade-offs — Only if a choice has non-obvious consequences
Use Mermaid diagrams for architecture and flow. Use tables for comparisons. Keep text precise and actionable.
Step 3 — Implement
Write core implementation code:
- Minimal, directly aligned with the design
- No unnecessary abstractions or speculative generalization
- No test cases, no test files
- Match vLLM codebase style (snake_case, type hints, docstrings only where non-obvious)
- Organize as: data structures → interfaces → core logic → integration points
Step 4 — Save Document
Save the complete design document as a Markdown file to ./outputs/ in the current working directory (create the directory if it doesn't exist). Filename: design-<feature-name>.md.
What ships with it
1 file 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 · 79 lines · 154 tokens per session scan A 3b6e1b1bb33a
vllm-feature-design is a skill published in the GitHub repository shen-shanshan/vllm-dev-skills (17 stars, last pushed 2d ago), licensed Apache-2.0. It adds 154 tokens to every session and 765 once invoked, about $0.0008 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
serving-llms-vllm
Serves LLMs with high throughput using vLLM's PagedAttention and continuous batching. Use when deploying production LLM APIs, optimizing inference latency/throughput, or serving models with limited GPU memory. Supports OpenAI-compatible endpoints, quantization (GPTQ/AWQ/FP8), and tensor parallelism.
outlines
Guarantee valid JSON/XML/code structure during generation, use Pydantic models for type-safe outputs, support local models (Transformers, vLLM), and maximize inference speed with Outlines - dottxt.ai's structured generation library.
human-agent-trust-exploit-detection
Detect social engineering, deceptive responses, false assurances, or prompts that induce unsafe user actions.
shap
Model interpretability and explainability using SHAP (SHapley Additive exPlanations). Use this skill when explaining machine learning model predictions, computing feature importance, generating SHAP plots (waterfall, beeswarm, bar, scatter, force, heatmap), debugging models, analyzing model bias or fairness, comparing…
glycobiology
Glycosylation site prediction and glycobiology analysis. N-glycosylation motif finding, O-glycosylation hotspot prediction, glycan structure resources. Lightweight, pure Python. For protein function queries use uniprot-database; for structure analysis use alphafold-database.
cellxgene-census
Query the CELLxGENE Census (61M+ cells) programmatically. Use when you need expression data across tissues, diseases, or cell types from the largest curated single-cell atlas. Best for population-scale queries, reference atlas comparisons. For analyzing your own data use scanpy or scvi-tools.