Skip to content
Agentmart
Submit a tool

steroid-research

This skill should be used after steroid-specify has written .memory/changes/<feature>/spec.md.

Skill by nzkbuild

IndexedMIT★ 3
Write a review

Chalk Labs

Need help deploying steroid-research?

Chalk Labs sets up agents like this inside real teams: connected to your tools, tested on your data, with a human in the loop where it matters.

Book a call

Install

Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.

npx agentmarket install skill/nzkbuild-steroid-research

Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.

Get the weekly diff

The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.

We use your email only for this newsletter. Privacy

About

This skill should be used after steroid-specify has written .memory/changes/<feature>/spec.md. It investigates the best tech stack, libraries, and patterns to implement the spec, writing findings to .memory/changes/<feature>/research.md.

Use cases: writing, research

No reviews yet. Installed it? Reviews come from verified installers only. How reviews work

Permissions

What this skill can do and what it asks for, from its SKILL.md. Checked Oct 25, 2026.

Tools
No tools found
Secrets
None asked for
Install scripts
None found

Network hosts it talks to 1

  • codeload.github.com · found in install command

Files it touches

  • ~/.claude/skills/steroid-research

What's inside

Skill name
steroid-research
Description
This skill should be used after steroid-specify has written .memory/changes/<feature>/spec.md. It investigates the best tech stack, libraries, and patterns to implement the spec, writing findings to .memory/changes/<feature>/research.md.
License
MIT
Instructions
399 lines in SKILL.md · view on GitHub

Files (1)

  • SKILL.md

SKILL.md (first lines)

# Steroid Research (Tech Investigation Agent)

This skill reads the spec from `.memory/changes/<feature>/spec.md` and investigates the best technical approach before the Architect commits to a plan.

## Governed Baseline

The live governed authority for this phase is:

- `Steroid runtime contract`
- `Steroid runtime contract`
- `Steroid runtime contract`
- `Steroid runtime contract`

In the live repo, this skill is the execution surface for the governed `steroid-research-system`.

The governed live research artifacts are:

- `.memory/changes/<feature>/context.md`
- `.memory/changes/<feature>/research.md`

Internal reference: `Steroid internal reference` (Steroid, MIT License) — simplified from 556 to ~150 lines for the Steroid non-technical context.

## When To Use

Activate immediately after `@steroid-specify` completes. Do not wait for user invocation. The spec.md file must already exist in `.memory/changes/<feature>/`.
Also read `.memory/changes/<feature>/context.md` as the governed internal codebase map from scan.

## Instructions

### 1. Phase Gate (Physical Enforcement)

Before doing anything, run the gate check:

```
node steroid-run.cjs gate research <feature>
```

If this command fails, STOP. The specification phase is not complete.
For UI-intensive work, this gate now auto-bootstraps `design-routing.json` and `design-system.md` when enough prompt/spec/vibe context exists.

### 2. No User Interaction

Do not ask the user any questions. Read `.memory/changes/<feature>/spec.md`, and if present `.memory/changes/<feature>/prompt.json`, investigate what's needed, and make all recommendations silently.

### 2. Research Philosophy

> Forked from Steroid: "Research is investigation, not confirmation."

- Treat your training data as **hypothesis, not fact**. Libraries change, APIs break, best practices evolve.
- **Verify before asserting** — don't state library capabilities without checking official docs or package registries.
- **Be honest about gaps** — "I couldn't verify this" is more valuable than a confident lie.
- **Be prescriptive, not exploratory** — "Use X" not "Consider X or Y". The Architect needs decisions, not options.

### 3. Circuit Breaker Mandate

If at any point you need to run a terminal command, you MUST use:
`node steroid-run.cjs '<command>'`
Direct terminal execution is strictly forbidden.

### 3b. Constraint Obedience (v5.5.0)

Read the `## Hard Constraints` section from spec.md. If the user explicitly requested a specific technology, version, or approach (e.g., "must use Framer Motion", "do NOT use a database"), the research MUST:

- Center around that constraint, not recommend alternatives
- Explain how to best use the constrained technology, not argue against it
- Only flag constraints that are technically impossible (e.g., "use React for an iOS native app")

If `prompt.json` exists, also preserve:

- `recommendedPipeline` as route context for how heavy the research should be
- `complexity` and `risk` as signals for how cautious the investigation should be
- `nonGoals` so you do not accidentally expand scope with optional tooling

**Design prep primitive:** If you need to refresh both UI artifacts manually in one shot, run:
`node steroid-run.cjs design-prep "<normalized prompt or user request>" --feature <feature> --write`

**Design routing receipt:** For UI-intensive work, the research gate now tries to auto-create this. If you need only the routing receipt, run:
`node steroid-run.cjs design-route "<normalized prompt or user request>" --feature <feature> --write`

If `design-routing.json` exists afterward, treat it as the routing receipt for which frontend source-library inputs Steroid should use during research.

**Design system artifact:** For UI-intensive work, the research gate now tries to auto-create this too. If you need only the design system artifact, run:
`node steroid-run.cjs design-system --feature <feature> --write`

`design-system.md` is not optional polish. It is the binding artifact that Architect and Engine now expect for UI-intensive work.

### 3c. Brownfield Context Check (v5.5.0)

Read vibe.md for the `Project Type` field. If it says **Brownfield**:

1. Run `node steroid-run.cjs scan` to detect the existing tech stack
2. Your research MUST respect the existing stack — do NOT recommend replacing it
3. Focus on libraries that integrate with what's already there
4. Note version compatibility with existing dependencies

If `prompt.json` says `continuationState` is `resume` or `mid-feature`, prefer incremental recommendations over broad re-architecture.

### 4. What To Investigate

For each user story in the spec, identify:

**Standard Stack** — What libraries/frameworks are the proven choice for this type of feature?

- Include specific version numbers
- Prefer stable, well-maintained packages over trendy new ones
- For a non-technical user's project, prefer zero-config or minimal-config tools

**Architecture Patterns** — What's the recommended project structure and design approach?

- File organization that scales
- State management approach
- Data flow patterns

**Don't Hand-Roll** — What problems look simple but have battle-tested solutions?

- Date/time handling → use a library
- Form validation → use a library
- Charting → use a library
- Never build what npm already solved

From nzkbuild/steroid-workflow, MIT.

More from nzkbuild/steroid-workflow

Signals

Stars
3
Forks
0
Open issues
0
Last commit (this folder)
Mar 26, 2026 (7 months ago)
Last release
Mar 16, 2026 (7 months ago)

From GitHub (fetched Oct 24, 2026).

Details

License
MIT
First seen
Oct 24, 2026
Last checked
Oct 24, 2026

More from nzkbuild

Done for you

Rather not set it up yourself?

Chalk Labs scopes, builds and runs agents for your team: the task audit, the eval harness, the integrations and the weekly ops report. Fixed scope, priced before we start.

See how we build AI agents →Read the case studies

Chalk Labs is a two-founder AI and growth studio. We build AI agents, CRMs and MVPs for Web3 and AI companies, and have shipped 50+ product and token launches.

launches shipped
50+
launches shipped
media placements
500+
media placements
founders on every account
2
founders on every account

Press we've placed: Forbes · CoinDesk · Decrypt · Cointelegraph · The Block

Rahil Jain, Co-founder, Chalk Labs. Ex-Goldman Sachs engineer. You talk to the person who builds it.

Get steroid-research running in your business

Tell us what you want to automate. A founder replies within one working day with a plan, not a pitch deck.

Goes to the Chalk Labs founders, not the newsletter. Privacy

Need help deploying steroid-research?

CallGet help
Cite this page

steroid-research skill. Agent Mart. https://agentmart.dev/skills/nzkbuild-steroid-research (last updated 2026-10-24)