AGENT SKILL

guixu

暂无简介

License 许可
Apache-2.0
最近更新
2026-11-14
Stars
★ 1.6k
Forks
491

03 SKILL.md 正文

<!-- Copyright (c) 2026 The State Key Laboratory of Blockchain and Data Security, Zhejiang University SPDX-License-Identifier: Apache-2.0 -->


name: guixu description: "Dataset discovery, valuation, and acquisition for AI training. Use when: (1) finding datasets for model training or fine-tuning, (2) evaluating dataset quality and relevance, (3) downloading datasets from Kaggle, HuggingFace, IPFS, or other sources, (4) assessing dataset licensing and provenance, (5) acquiring paid or free datasets. NOT for: generic web search (use browser), local file operations (use file tools), or data labeling/annotation tasks." metadata: { "openclaw": { "emoji": "?", "requires": { "bins": ["guixu"] }, "install": [ { "id": "brew", "kind": "brew", "formula": "guixu", "bins": ["guixu"], "label": "Install Guixu CLI (brew)", }, ], }, }

Guixu Data Agent

Guixu provides dataset discovery, valuation, and acquisition for AI agents. It searches across Kaggle, HuggingFace, IPFS, arXiv, DBLP, and other sources.

When to Use

✅ USE this skill when:

  • User asks to "find a dataset for training..."
  • User asks to "download a dataset about..."
  • User asks to "evaluate dataset quality..."
  • User asks about dataset licensing or provenance
  • User mentions specific datasets (Kaggle, HuggingFace, etc.)
  • User wants to search for training data

❌ DON'T use this skill when:

  • Generic web search → use browser or web search tools
  • Local file read/write → use file tools
  • Data labeling or annotation → use coding tools
  • Code implementation questions → use coding-agent skill

Workflow

Step 1: Parse Intent (REQUIRED first)

Always start with intent_parse to structure the request:

intent_parse(query="find me a cat image dataset for classification", task_type="classification")

This returns a structured QueryProfile with:

  • task_type: classification, detection, segmentation, etc.
  • keywords: dataset content terms
  • target_entity: main subject
  • data_standard: sample_unit, budget, schema expectations

Step 2: Search Datasets

Use dataset_search with keywords from intent_parse:

dataset_search(query="cat image", task_type="classification", limit=10)

Supported sources (leave empty to search all):

  • kaggle, huggingface, ipfs, bittorrent
  • arxiv, dblp, semanticscholar
  • defillama, rwa_xyz, guixu-hub
  • pansearch (cloud drives)

Filter options:

  • filters.max_price: maximum price in USD
  • filters.free_only: only free datasets
  • filters.license: specific license (e.g., "CC-BY-4.0")

Step 3: Evaluate Candidates

For each promising candidate, call dataset_evaluate:

dataset_evaluate(cid="kaggle:owner/dataset-name", task_description="cat image classification", task_type="classification", required_columns=["image_path", "label"])

This returns:

  • tcv_score: -100 (harmful) to +100 (highly valuable)
  • schema_fit: compatibility with required columns
  • community_signal: reviews and ratings

Step 4: Download

Once a dataset is selected:

dataset_download(cid="kaggle:owner/dataset-name")
# or
dataset_download(cid="hf:owner/dataset-name")

Free sources: uci:, openml:, zenodo:, figshare:, hf: (public), ipfs:, guixu-hub: Requires login: kaggle:

Tool Chaining Examples

Classification Dataset

1. intent_parse(query="find me a dog vs cat classification dataset", task_type="classification")
2. dataset_search(query="cat dog classification", task_type="classification", limit=10)
3. dataset_evaluate(cid="kaggle:username/dataset", task_description="dog cat binary classification", task_type="classification", required_columns=["image_path", "label"])
4. dataset_download(cid="kaggle:username/dataset")

Detection Dataset

1. intent_parse(query="find helmet detection dataset with bounding boxes", task_type="detection")
2. dataset_search(query="helmet detection bounding box", task_type="detection", limit=10)
3. For each candidate: dataset_evaluate(cid, task_description="helmet detection", task_type="detection", required_columns=["image_path", "bbox"])
4. Download best candidate

Tabular Finance Dataset

1. intent_parse(query="find stock price dataset for time series forecasting", task_type="forecasting")
2. dataset_search(query="stock price time series", task_type="forecasting", filters={source: "defillama"})
3. dataset_evaluate(cid, task_description="stock price prediction", task_type="forecasting", required_columns=["timestamp", "price"])
4. dataset_download(cid)

Important Notes

  • Always call intent_parse FIRST — it extracts task_type and keywords that improve search quality
  • Keywords should be CONTENT only — no task words like "classification" or "detection"
  • Check license before purchase — use require_license_review: true in evaluation
  • Budget enforcement — set filters.max_price or budget to limit spending
  • Free sources first — try guixuhub, huggingface, ipfs before paid sources

Error Handling

If dataset_search returns no results:

  • Try broader keywords
  • Remove source filters
  • Check if the source is spelled correctly

If dataset_evaluate fails:

  • Dataset may be unavailable
  • Try a different candidate

If dataset_download fails:

  • Check if login is required (Kaggle)
  • Verify the CID format is correct: source:owner/dataset

04 同仓库更多 Skills

hashgraph-online
hashgraph-online/awesome-codex-plugins

worktable-review-with-annotations

Create, inspect, reply to, update, or resolve situated Worktable annotations. Use when feedback belongs to a Doc section, block, text range, or HTML Doc, or when the user asks to address existing review feedback and coordinate a requested artifact edit.

★ 1.6k
hashgraph-online
hashgraph-online/awesome-codex-plugins

worktable-find-and-synthesize-context

Find, assess, and synthesize context from Worktable Docs and Records. Use when the user asks what a workspace contains, wants workspace-backed research, needs current facts reconciled across artifacts, or requests a cited synthesis without changing workspace state.

★ 1.6k
hashgraph-online
hashgraph-online/awesome-codex-plugins

worktable-create-or-update-docs

Create, edit, rename, or maintain narrative Worktable Docs. Use for plans, decisions, briefs, notes, research syntheses, and other durable prose when the user requests a new artifact or a safe update to an existing one.

★ 1.6k
hashgraph-online
hashgraph-online/awesome-codex-plugins

worktable-create-or-manage-records

Organize repeated structured information as durable Worktable Records. Use for research datasets, inventories, directories, pipelines, trackers, and other collections whose items need identity, fields, filtering, links, validation, or individual updates.

★ 1.6k
hashgraph-online
hashgraph-online/awesome-codex-plugins

worktable-create-or-manage-interactive-html

Create or improve durable visual and interactive artifacts in Worktable. Use for dashboards, trackers, calculators, mini apps, slide presentations, and other requests that benefit from a polished HTML Doc.

★ 1.6k
hashgraph-online
hashgraph-online/awesome-codex-plugins

worktable-collaborate-threads

Read, answer, and hand off messages through durable Worktable threads with users or other authenticated participants. Use when the user asks to check addressed messages, reply to a participant, continue a thread, request another agent's help, wait briefly for a response, or watch for comments and messages sent to the agent.

★ 1.6k