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/acoyfellow/unsurf/agents-mdgit clone --depth 1 https://github.com/acoyfellow/unsurfWrote 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/acoyfellow/unsurf/agents-md)<a href="https://agentmods.dev/instructions/acoyfellow/unsurf/agents-md"><img src="https://agentmods.dev/badge/instructions/acoyfellow/unsurf/agents-md.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 | $0.00610 | $0.00610 |
| Opus 5 | $0.00305 | $0.00305 |
| Sonnet 5 | $0.00122 | $0.00122 |
| Haiku 4.5 | $0.00061 | $0.00061 |
Grade A, and why
unsurf 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 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 — 82 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md — unsurf
What This Is
Turn any website into a typed API. Scout captures network traffic, infers schemas, outputs OpenAPI. Worker replays endpoints directly. Heal re-scouts when paths break.
- MCP Server: https://unsurf-api.coey.dev/mcp
- Docs: https://unsurf.coey.dev
- Directory: https://unsurf.coey.dev/directory
- NPM:
unsurf
Stack
- Effect — typed errors, DI, streams, retries
- Alchemy — infrastructure as TypeScript
- Drizzle — typed SQL (D1/SQLite)
- Cloudflare Workers — edge runtime
- Cloudflare Browser Rendering — headless Chrome
Architecture
src/
├── cf-worker.ts # Worker entry point
├── mcp.ts # MCP server entry point
├── cli.ts # CLI (unsurf search/lookup/publish)
├── domain/ # Effect Schema types
├── db/ # Drizzle schema + queries
├── services/ # Browser, Store, SchemaInferrer, OpenApiGenerator, Gallery, Directory
├── tools/ # Scout, Worker, Heal
└── ai/ # LLM Scout Agent (ScoutAgent, AnthropicProvider)
Directory
The API Directory is a community registry of unsurfed APIs at /directory.
Endpoints
GET /d/ # List all fingerprints
GET /d/:domain # Get domain info + capabilities
GET /d/:domain/:capability # Get endpoints for capability
GET /d/:domain/:method/:path # Get specific endpoint details
GET /d/:domain/spec # Get OpenAPI spec
POST /d/publish # Publish scouted API to directory
POST /d/validate # Validate before publishing
DELETE /d/:domain # Remove from directory
GET /search?q=... # Semantic search across directory
Seeding
Curated seed list at scripts/seed-apis.json. Run seeding script:
bun run scripts/seed.ts
Conventions
exactOptionalPropertyTypes: true- Biome for lint/format (tabs, 100 width)
- CI: check → docs → deploy
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 · 82 lines · 610 tokens per session scan A 5fc57d97fbe6
unsurf AGENTS.md is an instructions file published in the GitHub repository acoyfellow/unsurf (21 stars, last pushed 2mo ago), licensed MIT. It adds 610 tokens to every session, about $0.0030 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
scalar AGENTS.md
AGENTS.md instructions for scalar/scalar, covering agents.md - ai agent guide for scalar, project overview, prerequisites, first-time setup and commands.
sonicjs AGENTS.md
AGENTS.md instructions for SonicJs-Org/sonicjs, covering sonicjs agent guidelines, decision tree — code exploration, rtk (rust token killer), caveman mode and delegation — cavecrew subagents.
api-gateway CLAUDE.md
Claude Code instructions for membrane/api-gateway, covering claude.md, git & commit policy, working principles, 1. think before coding and 2. simplicity first.
LAP CLAUDE.md
Claude Code instructions for Lap-Platform/LAP, covering lap - lean api platform, what this is, quick reference, install and run tests.
apifable AGENTS.md
AGENTS.md instructions for ycs77/apifable, covering agents.md, keep this file strict, constraints, commands and rules.
robinhood-cli-mcp-api AGENTS.md
AGENTS.md instructions for zaydiscold/robinhood-cli-mcp-api, covering agents.md — driving robinhood-cli from an agent, what this is and how it works (read this for context), 0. tl;dr for an agent, 1. auth (browser-free, self-healing) and 2. discover the accounts — never hardcode them.