dunx: Skill for Claude Code

.claude/skills/spike/SKILL.md

spike is a skill for Claude Code from petarzarkov/dunx. It costs 74 tokens per session (645 once invoked), scanned A, original, MIT.

A measured investigation used to answer an open technical question before a design depends on it. The result is recorded as a verified constraint in the project architecture documentation.

In plain words
What is it for?
Testing questions about APIs, decorators, types, inheritance, or other runtime behaviour using temporary probes, then documenting the result.
Why use it?
It replaces assumptions about Bun or TypeScript behaviour with results from a real runtime and compiler check.

Skill for Claude Code

Written for Claude Code: installed under .claude/. Also seen: mentions subagents.

This is petarzarkov/dunx's own configuration. It tells Claude Code how to work on dunx itself, so it is not a mod to install elsewhere. Copy it as a starting point and replace the rules that are about this project. Everything dunx configures →

Reuse

Borrowing it

Nothing to install: this file belongs to petarzarkov/dunx. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.

Copy the file
curl -O https://raw.githubusercontent.com/petarzarkov/dunx/main/.claude/skills/spike/SKILL.md
Clone the repo
git clone --depth 1 https://github.com/petarzarkov/dunx

Made for: Claude Code.

Wrote 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.

agentmods badge for spike

README.md
[![agentmods](https://agentmods.dev/badge/skills/petarzarkov/dunx/spike.svg)](https://agentmods.dev/skills/petarzarkov/dunx/spike)
Your own site
<a href="https://agentmods.dev/skills/petarzarkov/dunx/spike"><img src="https://agentmods.dev/badge/skills/petarzarkov/dunx/spike.svg" alt="Measured on agentmods" height="20"></a>
Per session 74 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 645 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe. Third-party audits
  • NVIDIA SkillSpector pass 7 Sept 2026
How audits are shown
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5.1 $0.00074 $0.00645
Opus 5 $0.00037 $0.00322
Sonnet 5 $0.00015 $0.00129
Haiku 4.5 $0.00007 $0.00064

Measured 8d ago against content hash 0de155160343, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-08, from the pricing page.

Security

Grade A, and why

spike 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 8d 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.

.claude/skills/spike/SKILL.md · 49 lines

How it starts

The opening of the file, as written. The whole thing — 49 lines — stays where its author put it; the contents beside it link to each section on GitHub.

/spike

Every constraint in the Verified constraints section of docs/ARCHITECTURE.md was measured, not reasoned about. That is why the decisions above it hold. A spike keeps that property.

Procedure

  1. State the question as a falsifiable claim and name what it gates. From docs/architecture/constraints.md: "does @Post(path, { body: Schema }) constrain the method signature through the method decorator's generic?" gates Phase 3. "Does the WeakMap pending-drain survive subclassed controllers without Symbol.metadata?" gates Phase 2.
  2. Write a throwaway probe in the scratchpad directory, never under packages/ or examples/. It is not code that ships and it must not reach a commit, a coverage run, or a build.
  3. Run it on real Bun and record bun --version alongside the output.
  4. Delegate the probing when it is noisy or wide. Give a subagent the claim and the probe location; ask back for the literal command, the literal output, and a one-line verdict. Iterating a decorator probe through six type errors is exactly the kind of output that should never enter the main thread.
  5. Record the result in docs/ARCHITECTURE.md:
    • confirmed → a Verified constraints entry with the command, the Bun version, and the literal output in a fenced block. Match the existing terseness - the paramtypes: [ "Db", "Object", "Number" ] entry is the model.
    • refuted → write down the fallback and why, under the decision it affects. A rejected approach recorded is the whole point of that document.
    • Remove the item from Spikes to resolve. A resolved spike left listed is worse than no list.
  6. Delete the probe.

Rules

  • Never write "should work", "presumably", or "in theory" into docs/ARCHITECTURE.md. If it was not run, it does not go in.
  • A spike that changes the public API shape belongs before the code it gates, not after.
  • One spike, one claim. Two questions are two probes.
  • Probes may use anything - including the dialects this repo bans in shipped code (experimentalDecorators, reflect-metadata) if refuting them is the point. That is how the existing emitDecoratorMetadata entry was produced. Scratchpad only.

Read the full file on GitHub · 49 lines

Changes

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.

  1. 8d ago First seen · 49 lines · 74 tokens per session scan A 0de155160343

Subscribe to this mod's changes

spike is a skill published in the GitHub repository petarzarkov/dunx (111 stars, last pushed yesterday), licensed MIT. It adds 74 tokens to every session and 645 once invoked, about $0.0004 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.

Related

Other skills, from other repositories

nestjs

Use when building or structuring a NestJS backend — feature modules, providers and DI wiring, provider scopes and request-lifecycle order, where to bind guards/pipes/interceptors/filters, and testing with Test.createTestingModule. NOT a bare Express/Fastify service with no DI (that is nodejs), NOT framework-agnostic…

ericrisco/rsc-harness · 81 tokens

bun

Use this skill when building with Bun runtime — Bun APIs, testing, package management, shell scripting, hot reload, SQLite, file I/O. This skill enforces: built-in APIs over npm equivalents, Bun test runner, bun install speed optimization, Bun.shell for scripts. Do NOT use for: Node.js-specific APIs…

j4flmao/agent-skills · 80 tokens

nifra

Use when writing, reviewing, or debugging code in a Nifra project (@nifrajs/ packages, nifra CLI, server()/defineContract, loaders and actions, file routes under routes/). Explains how to reach Nifra's live MCP tools so signatures come from the installed version instead of memory, and which sibling skill to load for…

nifrajs/nifra · 83 tokens

express-prisma-pattern

House pattern for Express + Prisma backends structured as manual Clean Architecture (controller → usecase → repository, wired by hand in a factory) — the real shape of dream-book-api and simple-management-api. Use whenever creating a new route, usecase or domain in an Express backend on this stack, even if the request…

JoaoEquer/Oficina · 85 tokens

api-errors

McpError constructor, JsonRpcErrorCode reference, and error handling patterns for @cyanheads/mcp-ts-core. Use when looking up error codes, understanding where errors should be thrown vs. caught, or using ErrorHandler.tryCatch in services.

cyanheads/federal-reserve-mcp-server · 54 tokens

Node Expert

Node Expert delivers production-grade technical work with clear architecture, tests, maintainability, and operational awareness.

ahmadrrrtx/xr · 23 tokens