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/nifrajs/nifra/agents-mdgit clone --depth 1 https://github.com/nifrajs/nifraWrote 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/nifrajs/nifra/agents-md)<a href="https://agentmods.dev/instructions/nifrajs/nifra/agents-md"><img src="https://agentmods.dev/badge/instructions/nifrajs/nifra/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 | $0.02352 | $0.02352 |
| Opus 5 | $0.01176 | $0.01176 |
| Sonnet 5 | $0.00470 | $0.00470 |
| Haiku 4.5 | $0.00235 | $0.00235 |
Grade A, and why
nifra AGENTS.md scanned grade A with 1 finding 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 4d 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.
Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
`env` is forwarded from `app.fetch(request, { env })` (Workers bindings). Validate at the boundary. How it starts
The opening of the file, as written. The whole thing — 157 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md - Nifra quick reference for coding agents
Nifra is a Bun-native, multi-runtime, end-to-end-typed full-stack TS framework. This is the
copy-paste cookbook so you don't have to read the source. Deeper docs live in docs/ and at
/docs/contract. Machine surfaces for agents: nifra check --json (the done-gate),
nifra context, and the MCP server (nifra mcp).
1 · Define a route (typed, chainable)
import { server } from "@nifrajs/core/server"
export const app = server()
.get("/users/:id", (c) => ({ id: c.params.id })) // c.params typed from the path literal
.get("/files/*path", (c) => ({ path: c.params.path })) // trailing wildcard
Return a value → serialized to JSON. Return a Response → used as-is.
2 · Validate a body/query at the trust boundary (auto-422)
import { t } from "@nifrajs/schema"
app.post(
"/users",
{ body: t.object({ name: t.string({ minLength: 1 }), age: t.number() }) },
(c) => ({ id: crypto.randomUUID(), name: c.body.name }), // c.body is typed + already validated
)
t.object rejects unknown fields by default (additionalProperties: false) → a structured 422 with { path: [...] } issues
before the handler runs. Need to allow extras? t.looseObject. Query schema: { query: t.object({…}) } → c.query.
3 · Read platform env (KV / D1 / secrets)
// Type it ONCE on server<Env>() → c.env is typed everywhere below (no per-binding cast).
const app = server<{ MY_KV: KVNamespace; API_SECRET: string }>().get(
"/x",
(c) => c.env.API_SECRET, // c.env: Env (it's `unknown` if you don't pass <Env>)
)
env is forwarded from app.fetch(request, { env }) (Workers bindings). Validate at the boundary.
4 · Set status / headers, and the control-flow rule
app.post("/things", { body: t.object({ name: t.string() }) }, (c) => {
c.set.status = 201
c.set.headers["x-created"] = "1"
return { ok: true }
})
For full control, return a Response. Throw rule: throw new Response("", { status: 404 }) =
an intentional HTTP response (control flow - bypasses _error); throw new Error(…) = the nearest
_error boundary / a 500.
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.
- 4d ago First seen · 157 lines · 2,352 tokens per session scan A fd2895d9edc4
nifra AGENTS.md is an instructions file published in the GitHub repository nifrajs/nifra (2 stars, last pushed yesterday), licensed MIT. It adds 2,352 tokens to every session, about $0.0118 per session on Opus 5. A static security scan graded it A with 1 finding (makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other instructions, from other repositories
keryx CLAUDE.md
Instructions for actionhero/keryx, covering claude.md, project overview, monorepo structure, development environment and environment setup.
dawnai AGENTS.md
AGENTS.md instructions for cacheplane/dawnai, covering agents.md, what this is (and isn't), workspace map, definition of done and conventions.
veryfront-code AGENTS.md
AGENTS.md instructions for veryfront/veryfront-code, covering veryfront code agent guide, default working style, commands, command discovery and mcp server.
sxo AGENTS.md
Instructions for gc-victor/sxo, covering agents.md, quick start (tl;dr), special error pages (404/500), info and purpose.
stitchkit AGENTS.md
Instructions for max-listov/stitchkit, covering stitchkit — agent guide, two roads — pick yours, rules, stack and commands.
stitchkit CLAUDE.md
Instructions for max-listov/stitchkit: This project's agent guide is AGENTS.md — the tool-agnostic standard also read by Cursor, Codex and other tools. It is imported below so Claude Code picks it up too.