iowarp/clio-agent-marketplace/audit-dataset

Trust tier 3
  • community
Source check pending

Checking that the SKILL.md file is still in the repository.

Inspect one dataset in a child, adapt checks to its real files and the parent's question, and return grounded findings, source references, optional outputs and unresolved scientific questions.

Install

npx vskill@latest install iowarp/clio-agent-marketplace/audit-dataset

SKILL.md

You are already running this skill in the delegated child. Complete the assigned inspection directly; do not recursively delegate it.

The assignment must identify the source bundle and should identify the question or audit scope. If the source is unreadable, report the actual access problem. Treat it as read-only input even when writes are technically permitted. Save notes, code and outputs outside it, usually in the Agent workspace state at <workspace_state>/datasets/<key>/, resolved from CLIO_AGENT_WORKSPACE_STATE_DIR, or another permitted output directory.

  1. Load onboard-dataset and inspect the files needed for the assignment. For

phenotyping, use phenotyping-onboarding-checks for scientific checks. Scripts, cards and schemas are optional aids. Adapt or replace them when assumptions do not fit; an unfamiliar export version is not a blanket blocker.

  1. Keep uncertain treatment meanings, units, calibration and identity mappings

explicit. Return these questions to the parent rather than guessing or changing source data. Continue checks that do not depend on their answers.

  1. Check calculations and outputs against source observations, counts and join

behavior. Process completion or a passing example schema alone does not establish scientific validity. Do not convert a whole dataset merely to fill a required set of tables. The parent owns final presentation.

  1. Return a concise account of inspected scope, evidence, findings, limitations,

questions, paths of any notes/code/outputs, and checks with their results. For reusable code, include its invocation and dependencies. If no loader or views were needed, say so. Identify any specific unfinished work.

The parent checks the evidence supporting its answer: inspecting a sample, repeating a calculation, comparing counts or rerunning selected code when warranted. A mandatory second execution of the entire pipeline is not required. A partial result must be labelled accurately. Continue until the assigned scope is satisfied or a concrete access or interpretation dependency remains.

Security

  • Security scanPassed, 90 of 100Pattern checks on the skill's files.
  • Model reviewPassed

Quality evaluation

  • Task evaluationNot runNo evaluation yet. Run one in Skill Studio.

Source

Popularity

7-day trend score
5

Not refreshed yet.

Works with

Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.

  • Claude Code
  • Cursor
  • GitHub Copilot
  • Windsurf
  • Codex
  • Gemini CLI
  • Cline
  • Amp
  • Roo Code
  • Goose
  • Continue
  • Kilo Code
  • Junie
  • Trae
  • OpenHands
  • Qwen Code
  • AI Agent
  • Tabnine Chat
  • Sourcegraph Cody
  • Amazon Q
  • Firebase Genkit
  • Sweep
  • Devin
  • Phind
  • Mentat
  • Claude Desktop
  • Aider
  • GPT Pilot
  • Plandex
  • bolt.new
  • v0
  • Replit Agent
  • Anse
  • TypingMind
  • BetterChatGPT
  • LibreChat
  • Warp AI
  • Pieces
  • Fig AI
Embed badge
vskill VERIFIED badge
[![vskill VERIFIED](https://verified-skill.com/api/v1/skills/iowarp/clio-agent-marketplace/audit-dataset/badge)](https://verified-skill.com/skills/iowarp/clio-agent-marketplace/audit-dataset)

Something wrong with this skill or its page? Tell us.