Atlas / Skills / dicklesworthstone / Minimax Web Search

Minimax Web SearchSAFE

skills/dicklesworthstone/minimax-web-search

High-performance AI coding agent CLI written in Rust with zero unsafe code

Verdict
SAFE
Grade
B
Trust score
89 /100
Version
—
Hosts
1 documented
License
NOASSERTION
Stars
1,872
01

Overview

High-performance AI coding agent CLI written in Rust with zero unsafe code

Read from source at commit 46a2537504c9OBSERVED · 2026-10-10
02

Host compatibility

What the documentation claims. We have not run a compatibility test.

HostStatusNotes
claude-codementioned
03

What it tells the agent

The instruction file, verbatim from the audited commit — this is the text the model reads, and the surface the audit's instruction layer examines. Quoted here so you can judge it without cloning anything.

---
name: minimax-web-search
description: Search the web for current information using the web_search tool
---

# MiniMax Web Search Skill

Use this skill when you need to find current, real-time, or authoritative information that may have changed since my knowledge cutoff.

## How to Use

Call the `web_search` tool directly with a query:

```
web_search({ query: "your search query" })
```

## When to Use

Use `web_search` when:

- **Current events**: News, recent developments, latest releases
- **Fact verification**: Checking if information is still accurate
- **Technical updates**: New versions, patches, breaking changes
- **Dynamic content**: Prices, availability, schedules
- **Authoritative sources**: Official documentation, API references

## When NOT to Use

Do NOT use `web_search` when:

- **General knowledge** I likely know (historical facts, basic concepts)
- **Coding help** for common patterns (unless asking about latest practices)
- **Opinion-based questions** without specific recent information needs
- **Files already in context** or accessible via `read` tool

## Usage

```
web_search({
  query: "your search query here"
})
```

## API Details

**Endpoint**: `POST {api_host}/v1/coding_plan/search`

**Request Body**:
```json
{
  "q": "your search query"
}
```

**Response Format**:
```json
{
  "organic": [
    {
      "title": "Result Title",
      "link": "https://example.com",
      "snippet": "Brief description...",
      "date": "2025-01-15"
    }
  ],
  "related_searches": [
    {"query": "related query 1"},
    {"query": "related query 2"}
  ],
  "base_resp": {
    "status_code": 0,
    "status_msg": "success"
  }
}
```

## Tips for Better Results

1. **Be specific**: "TypeScript 5.4 new features" vs "TypeScript updates"
2. **Include context**: "React 19 server components vs client components"
3. **Use recent keywords**: "2024" or "2025" when timing matters
4. **Target authoritative sources**: Include product names, official sites
5. **Use 3-5 keywords** for optimal results

## Examples

### Good Queries
- "Node.js 22 new features"
- "Claude Code vs pi coding agent 2025"
- "React 19 release date official"
- "Docker desktop alternatives 2024"
- "TypeScript 5.4 generics tutorial"

### Poor Queries
- "how to use TypeScript" (too general)
- "is Python popular" (opinion, not time-sensitive)
- "help me debug my code" (not a search task)
- "best programming language" (opinion-based)

## Response Format

Results include:
- **Titles** and **URLs** for each result
- **Snippets** with relevant context
- **Dates** when available
- **Related searches** for follow-up queries

You can use the suggestions to refine your search or ask the user which direction they'd like to explore.

## Error Handling

- **Status code 1004**: Authentication error - check API key and region
- **Status code 2038**: Real-name verification required
- **Other errors**: Check Trace-Id in error message for support
04

Trust audit

SAFEgrade B · trust 89/100 Nothing in the source contradicts what it says it does. Grade A is reserved for packages that have also passed the behavioural sandbox.

LayerWhat it checksResult
L0Provenance & inventoryWARN
L1Static analysis of the codePASS
L2Instruction surface (what it tells the agent)PASS
L3Class-specific surfacePASS
L4Behavioural (sandbox)SKIPPED

What the source does

Filesystem
none-observed
Network
none-observed
Shell
none-observed
Dependencies
pinned
Secrets in source
none-found

Findings (1)

MEDIUMInventory / provenance · inv.symlink · CWE-1104
.beads/beads.jsonl
.beads/beads.jsonl
Why it matters. link not followed

Gates applied: no_behavioural_pass.

Audited 2026-10-10 · audit v0.4.1 · source sha 46a2537504c9full audit observations/trust-audit/skill/dicklesworthstone__minimax-web-search.json · Report an issue / request a re-scan
05

Audit history

Every audit this skill has had.

DateSourceVerdictGradeScoreChange
2026-10-1046a2537504c9SAFEB89first audit
06

Questions

What does the Minimax Web Search skill do?

High-performance AI coding agent CLI written in Rust with zero unsafe code

Is Minimax Web Search safe to install?

The audit found nothing in the source that contradicts what it says it does, and graded it B (89/100). Grade A is held back for packages that have also passed a sandboxed behavioural run, which is why a clean skill reads B.

What can Minimax Web Search access on my machine?

The audit observed no filesystem, network or shell use at all in its source.

Which assistants does Minimax Web Search work with?

Its documentation mentions claude-code. That is what the text claims, not a compatibility test we ran.

How current is this page?

The grade is for one exact copy of the source (46a2537504c9), read on 2026-10-10. The repository is watched, and a new audit runs when it changes — this is the first audit.

Advertisement