# Agentwheel Smoke

> Run a read-only smoke diagnostic for the current AI harness and its Agentwheel-managed installation state. Use when asked to smoke test an agent, verify fleet deployment, inspect whether a harness is installed in the expected Agentwheel path, or summarize local Agentwheel status without mutating files or external systems.

- **Type:** Skill
- **Install:** `agentstack add skill-nestdevlab-agentwheel-agentwheel-smoke`
- **Verified:** Yes — security-reviewed for prompt injection and unsafe behavior
- **Seller:** [NestDevLab](https://agentstack.voostack.com/s/nestdevlab)
- **Installs:** 0
- **Category:** [Agent Skills](https://agentstack.voostack.com/c/agent-skills)
- **Latest version:** 0.1.0
- **License:** MIT
- **Upstream author:** [NestDevLab](https://github.com/NestDevLab)
- **Source:** https://github.com/NestDevLab/agentwheel/tree/main/skills/agentwheel-smoke
- **Website:** http://www.nestdev.it/agentwheel/

## Install

```sh
agentstack add skill-nestdevlab-agentwheel-agentwheel-smoke
```

Requires the [AgentStack CLI](https://agentstack.voostack.com/docs/cli). Works with Claude Code, Cursor, and any MCP-compatible agent.

## About

# Agentwheel Smoke

Use this skill to answer: "What harness am I running in, where was this skill installed, and what does Agentwheel think about this target?"

The default workflow is read-only. Do not run `agentwheel install`, `agentwheel update`, `agentwheel uninstall`, `agentwheel remember`, `agentwheel eject`, or any file-writing fix while using this skill unless the user separately approves that mutation scope.

## Quick Start

From the user's current workspace, run the bundled smoke script:

```bash
node path/to/agentwheel-smoke/scripts/smoke.mjs
```

Use JSON output only when another tool or script needs to consume the result:

```bash
node path/to/agentwheel-smoke/scripts/smoke.mjs --json
```

When the user names an Agentwheel target, pass only read-only status target flags:

```bash
node path/to/agentwheel-smoke/scripts/smoke.mjs --agent lab-codex
node path/to/agentwheel-smoke/scripts/smoke.mjs --profile daily
node path/to/agentwheel-smoke/scripts/smoke.mjs --all
```

## Report The Result

Summarize the script output in four parts:

- Overall: `PASS`, `WARN`, or `FAIL`.
- Harness: detected adapter, installation type, runtime root, and skill path.
- Agentwheel: CLI availability, version, config locations, packages, agents, profiles, and status result.
- Recommended next command: prefer `agentwheel status`, `agentwheel plan`, or `agentwheel install --dry-run`.

Treat these as important signals:

- `PASS`: skill path maps to a documented Agentwheel harness target and `agentwheel status` reports no pending work.
- `WARN`: the skill is running from package source, status is missing, config is absent, manifest is missing, or pending work is reported.
- `FAIL`: Agentwheel CLI is unavailable, the skill path maps to a known wrong target, `agentwheel status` exits unsuccessfully, or drift/conflict appears.

## Compatibility Checks

Use these documented skill targets when explaining path compatibility:

- Codex local: `.agents/skills//SKILL.md`; Codex user: `~/.agents/skills//SKILL.md`.
- Claude local: `.claude/skills//SKILL.md`; Claude user: `~/.claude/skills//SKILL.md`.
- Copilot local: `.github/skills//SKILL.md`; Copilot user: `~/.copilot/skills//SKILL.md`.
- OpenClaw local: `skills//SKILL.md`; OpenClaw user: `~/.openclaw/skills//SKILL.md`.
- Hermes user: `~/.hermes/skills//SKILL.md`; local Hermes skills are unsupported by default.

Flag `.codex/skills//SKILL.md` as incompatible for Agentwheel-managed Codex skills.

## Safety

This skill is diagnostic. It may inspect files and run read-only Agentwheel commands, but it must not mutate runtime outputs, package source, Google Workspace data, Git state, registry state, or deployment state.

## Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

- **Author:** [NestDevLab](https://github.com/NestDevLab)
- **Source:** [NestDevLab/agentwheel](https://github.com/NestDevLab/agentwheel)
- **License:** MIT
- **Homepage:** http://www.nestdev.it/agentwheel/

Install and usage instructions live in the source repository linked above.

## Pricing

- **Free** — Free

## Security capabilities

Automated source analysis of v0.1.0 — what this tool can access:

- **Network access:** no
- **Filesystem access:** no
- **Shell / process execution:** no
- **Environment & secrets:** no
- **Dynamic code execution:** no

*"Yes" means the capability is present in the source — more access means more to trust, not that it is unsafe.*


## Versions

- **0.1.0** — security scan: passed — Imported from the upstream source.

## Links

- Listing page: https://agentstack.voostack.com/l/skill-nestdevlab-agentwheel-agentwheel-smoke
- Seller: https://agentstack.voostack.com/s/nestdevlab
- Browse the marketplace: https://agentstack.voostack.com/browse

---
Listed on AgentStack — the marketplace for AI agent skills and MCP servers. Every listing is security-reviewed. Creators keep 70%.