Borrowing it
Nothing to install: this file belongs to snyk-labs/snyk-vulnbench. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/snyk-labs/snyk-vulnbench/main/.agents/skills/benchmark-report-writer/SKILL.mdgit clone --depth 1 https://github.com/snyk-labs/snyk-vulnbenchWrote 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/snyk-labs/snyk-vulnbench/benchmark-report-writer)<a href="https://agentmods.dev/skills/snyk-labs/snyk-vulnbench/benchmark-report-writer"><img src="https://agentmods.dev/badge/skills/snyk-labs/snyk-vulnbench/benchmark-report-writer/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/snyk-labs/snyk-vulnbench/benchmark-report-writer"><img src="https://agentmods.dev/badge/skills/snyk-labs/snyk-vulnbench/benchmark-report-writer.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 | $0.00184 | $0.05973 |
| Opus 5.5 | $0.00074 | $0.02389 |
| Sonnet 5.5 | $0.00037 | $0.01195 |
| Haiku 4.5 | $0.00018 | $0.00597 |
Grade A, and why
benchmark-report-writer 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 27d 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 — 324 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Benchmark Report Writer
Instructions
Produce a publishable markdown benchmark report from eval results and, when available, a benchmark guide. The report should match what a reader expects from frontier AI lab writeups: a short summary that sets honest expectations, a methodology section that earns the reader's trust, a results section anchored by a leaderboard and optional generated visuals, a qualitative analysis section that explains what the numbers mean, and a limitations section that doesn't flinch.
The goal is a report that a reader can skim in two minutes or read thoroughly in twenty, and that stands up to scrutiny because every claim is traceable to the data.
For security benchmarks where the tool under discussion also defines the reference set, do not lead with "Tool X scored 100%." That framing looks like grading your own homework. Lead with the most novel, externally useful finding the data supports: repeatability across identical runs, complementarity between model review and SAST, or a cost-quality inversion. Treat reference-set F1 as context, not the headline.
For VulnBench 2.0 attacker-reachable results, lead with Attacker-Reachable
Vulnerability Recall (primaryMetric: "attacker-reachable-vulnerability-recall"). Precision and F1 remain secondary
quality/noise guardrails. Never blend V1 F1, V2 recall, and fix rate into one
headline; use config-aggregate.byGroundTruth whenever top-level quality fields
are null.
Use version-first release naming in prose: VulnBench 1.0 and VulnBench
2.0 (or VulnBench 1/VulnBench 2 after first mention). The published formal
title Snyk VulnBench JS 1.0 and descriptive variant Snyk VulnBench Apps
2.0 are valid on first mention, but do not require the JS or Apps
qualifier in subsequent references. Never rename historical paper titles,
release URLs, or source-snapshot directories.
For public reports, avoid internal fixture codenames as prose labels. Use exact fixture
IDs in inline code when a specific fixture must be named, such as
js-project-nightowl or js-project-tigerteam. For chart titles, captions, section
headings, and narrative labels, prefer generic descriptions such as "larger multi-file
fixture" or "SQL-shaped mock helper fixture" unless the chart is explicitly a per-task
lookup table where the fixture name helps the reader map back to the data.
What ships with it
5 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.
- 27d ago Changed · +27 lines ebf867d1fb07
- 1mo ago First seen · 297 lines · 184 tokens per session scan A f6f951d0e8cb
benchmark-report-writer is a skill published in the GitHub repository snyk-labs/snyk-vulnbench (4 stars, last pushed 14d ago), licensed Apache-2.0. It adds 184 tokens to every session and 5,973 once invoked, about $0.0007 per session on Opus 5.5. 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
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…
systematic-debugging
Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes.
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…
chat-pet-sprite-creation
Use when creating or changing VS Code chat pet sprite art, sprite sheets, state animations, eye treatments, Stable/Insiders variants, or pet transitions under src/vs/workbench/contrib/chat/browser/widget/media/chatPet.
cpu-profile-analysis
Analyze V8/Chrome CPU profiles (.cpuprofile) and DevTools trace files (Trace-.json). Use when: profiling performance, investigating slow functions, comparing code paths, finding bottlenecks, analyzing timeToRequest, understanding call trees from sampling profiler data, analyzing layout/paint/rendering, investigating…
insight-error-page
Write or audit an insight-kind error page for the Next.js dev overlay. Use when creating a new errors/ .mdx page, auditing an existing one, or checking that a page matches the framework fix cards. Covers page structure, title alignment, FixCard cards with Copy prompt button, code snippets, terminology verification…