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.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/instructions/snyk-labs/snyk-vulnbench/agents-md)<a href="https://agentmods.dev/instructions/snyk-labs/snyk-vulnbench/agents-md"><img src="https://agentmods.dev/badge/instructions/snyk-labs/snyk-vulnbench/agents-md/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/instructions/snyk-labs/snyk-vulnbench/agents-md"><img src="https://agentmods.dev/badge/instructions/snyk-labs/snyk-vulnbench/agents-md.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 | $0.03690 | $0.03690 |
| Opus 5.5 | $0.01476 | $0.01476 |
| Sonnet 5.5 | $0.00738 | $0.00738 |
| Haiku 4.5 | $0.00369 | $0.00369 |
Grade A, and why
snyk-vulnbench AGENTS.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 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 — 216 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Coding Agent Security Benchmark
Project Purpose
Benchmarking framework that runs AI coding agents (primarily Claude Code via the TypeScript Agent SDK) against standardized security tasks, collecting metrics to compare:
- Different models (claude-opus-4-6 vs claude-sonnet-4-6 vs claude-haiku-4-5)
- Different MCP configurations (with/without security tools like Snyk, semgrep, etc.)
- Different system prompts and agent configs
Primary Eval Categories
- VulnBench 1.0 find categories (
find-vulns,llm-find-vulns,app-find-vulns): Given a codebase with known vulnerabilities infindings.json, how many can the agent correctly identify using the original type-only scorer? - VulnBench 2.0 attacker-reachable findings (
attacker-reachable-find-vulns): Compare reported source-to-sink flows against human-curatedfindings-attacker-reachable.jsonground truth. - fix-vulns: Given a codebase with known vulnerabilities, how many can the agent correctly fix?
Tech Stack
- Runtime: TypeScript + Node 24, pnpm
- Agent SDK:
@anthropic-ai/claude-agent-sdk(TypeScript) — wraps Claude Code CLI - Anthropic SDK:
@anthropic-ai/sdk— for token counting and direct API calls (scoring, judging) - Codex Security: pinned
@openai/codex-securityCLI — dedicated full-repository security harness - Claude Code CLI: available at
/home/node/.local/bin/claude
Architecture
src/
types.ts # Core interfaces + EVAL_CATEGORIES constant
runner.ts # Agent SDK wrapper — runs a task and collects metrics
runners/ # Claude, Codex, Codex Security, DeepSec, and command adapters
scorer.ts # Scoring logic per eval category
reporter.ts # Output to console table + JSONL
evals/
loader.ts # Scans evals/tasks/*.json + evals/run-configs.json at startup
index.ts # CLI entry point
evals/
tasks/ # One JSON file per eval task — add a file to add a new task
js-find-vulns.json
js-fix-vulns.json
python-find-vulns.json
run-configs.json # Array of RunConfig objects — edit to add/change model configs
fixtures/
js-project-tigerteam/ # Each fixture is a self-contained directory
project/ # Agent's working directory (source code)
app.js
findings.json # Ground-truth metadata (outside agent cwd)
python-project-cobalt/
project/
app.py
findings.json
app-project-keystonebank/
project/ # Shared application source
findings.json # VulnBench 1.0 ground truth
findings-attacker-reachable.json # VulnBench 2.0 ground truth
results/ # Ignored runtime execution bundles and JSONL snapshots
publications/ # Versioned, checksummed completed benchmark archives
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 · +51 lines · +1,172 tokens per session 3cc58c52e787
- 1mo ago First seen · 165 lines · 2,518 tokens per session scan A 6ad3cf7202f2
snyk-vulnbench AGENTS.md is an instructions file published in the GitHub repository snyk-labs/snyk-vulnbench (4 stars, last pushed 14d ago), licensed Apache-2.0. It adds 3,690 tokens to every session, about $0.0148 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 instructions, from other repositories
vscode buildNext.instructions.md
Working notes and architecture documentation for the new esbuild-based build system in build/next. Use when making changes to the new build pipeline (transpile/bundle commands, NLS plugin, source-map handling, resource copying, or self-hosting watch tasks).
next.js AGENTS.md
AGENTS.md instructions for vercel/next.js, covering next.js development guide, codebase structure, monorepo overview, core package: packages/next and other important packages.
vscode oss-third-party-notices.instructions.md
Instructions for microsoft/vscode, covering vs code oss third-party-notices pipeline, architecture, pipeline flow in ci, cache eligibility and rollout and applying the notice (cutover).
langchain AGENTS.md
AGENTS.md instructions for langchain-ai/langchain, covering global development guidelines for the langchain monorepo, corridor security analysis, project architecture and context, monorepo structure and development tools & commands.
vscode copilot-instructions.md
Copilot instructions for microsoft/vscode, covering vs code copilot instructions, project overview, root folders, core architecture (src/ folder) and built-in extensions (extensions/ folder).
vscode model-prompts.instructions.md
Model-specific prompt authoring and registry guidelines.