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 agents/alex-alecu/garden-desk/exploregit clone --depth 1 https://github.com/alex-alecu/garden-deskWrote 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/agents/alex-alecu/garden-desk/explore)<a href="https://agentmods.dev/agents/alex-alecu/garden-desk/explore"><img src="https://agentmods.dev/badge/agents/alex-alecu/garden-desk/explore.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 | $0.00035 | $0.00154 |
| Opus 5 | $0.00017 | $0.00077 |
| Sonnet 5 | $0.00007 | $0.00031 |
| Haiku 4.5 | $0.00003 | $0.00015 |
Grade A, and why
explore 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.
What it actually says
Explore Agent
Investigate the assigned question without changing files or external state. Start with the most relevant local instructions, contracts, code, and tests; follow references only when they clarify the decision.
Return the answer first, then the concrete evidence: relevant paths, behavior, constraints, and uncertainties. Distinguish observed facts from inferences. Keep the handoff self-contained so the parent can act without relying on hidden context.
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 · 15 lines · 35 tokens per session scan A 3d053c0e83e6
explore is an agent published in the GitHub repository alex-alecu/garden-desk (6 stars, last pushed yesterday), licensed Apache-2.0. It adds 35 tokens to every session and 154 once invoked, about $0.0002 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 2026-08-31.
Other agents, from other repositories
matrix
This page renders the Tier-1 agent × model-family integration matrix truthfully from the authoritative test suite in tests/integrations/ — the matrix cells (testagentsmatrix.py), the family aliases and strict-xfail rules (conftest.py), and the pilot run recorded in tests/integrations/README.md.
claude-code
Point Anthropic's Claude Code at a local rapid-mlx server. Claude Code speaks the Anthropic Messages API (POST /v1/messages); rapid-mlx implements that route natively, so you can drive Claude Code with any local model.
openhands
Point OpenHands (formerly OpenDevin) at a local rapid-mlx server. OpenHands drives its CodeActAgent inside a Docker sandbox and reaches the model over the OpenAI-compatible chat completions API (POST /v1/chat/completions) via LiteLLM.
qwen-code
Point Qwen Code at a local rapid-mlx server. Qwen Code is Alibaba's gemini-cli fork tuned for Qwen tool-calling; it speaks the OpenAI-compatible chat completions API (POST /v1/chat/completions) via an OpenAI entry in modelProviders that maps 1:1 onto rapid-mlx's default endpoint.
hermes-agent
Point Nous Research's Hermes Agent at a local rapid-mlx server. Hermes is a tool-heavy CLI agent (it injects up to 62 tools per request) that speaks the OpenAI-compatible chat completions API (POST /v1/chat/completions).
opencode
Point OpenCode at a local rapid-mlx server. OpenCode is a Claude-Code-like terminal coding agent that speaks the OpenAI-compatible chat completions API (POST /v1/chat/completions) via the @ai-sdk/openai-compatible provider.