sergeinikolenko/clearml-codex-plugin/clearml

Claude Code and agent skill by sergeinikolenko/clearml-codex-plugin

Trust tier 3
  • community
Source check pending

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

Use ClearML MCP tools to inspect experiments, compare runs at equal progress, summarize projects, and open or return ClearML UI links when requested.

Install

npx vskill@latest install sergeinikolenko/clearml-codex-plugin/clearml

SKILL.md

ClearML

Use this skill when the user asks about ClearML experiments, metrics, runs, logs, task status, project snapshots, or ClearML comparison links.

Defaults

  • Prefer read-only MCP tools from this plugin. Never create, enqueue, archive, delete, or mutate ClearML objects.
  • Do not ask the user to paste credentials. The plugin reads the standard ClearML SDK configuration or environment variables.
  • When comparing training runs, compare at a shared step by default. Latest-only comparisons can be misleading when tasks have different progress.
  • Lead with task status and domain metrics that answer the user's question, then mention losses/logs as supporting evidence.
  • Prefer outcome metrics over loss curves when the user asks for a domain verdict.
  • If task ids are not supplied, search tasks first by project, name, tags, and status.

Tool Routing

  • Use clearml_search_tasks to find task ids by project/name/tags/status.
  • Use clearml_project_snapshot for questions like "what is running in this project" or "show the latest runs".
  • Use clearml_experiment_slice when the user wants a compact data dump for several experiments.
  • Use clearml_compare_tasks for focused equal-step metric tables.
  • Use clearml_scalar_series for one metric over time.
  • Use clearml_task_logs and clearml_task_events when the user asks whether a run is healthy or stuck.
  • Use clearml_compare_url when the user asks for a ClearML UI comparison link.

Browser Behavior

If the user asks to open or inspect a ClearML page, first build or identify the URL. If Browser tools are available in the thread, open the URL in the in-app browser and inspect it. If Browser tools are not available, return the URL and say that a browser-enabled thread can inspect the page.

Output Shape

For comparisons, report:

  1. Compared task aliases and ids.
  2. Shared target step and how it was selected.
  3. Direct verdict on the requested metrics.
  4. Caveats about missing metrics, different metric variants, or very early steps.
  5. ClearML compare URL when project id is known.

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

Repository stars
1
7-day trend score
3

Stars and forks are for the whole GitHub repository, not this skill alone. 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/sergeinikolenko/clearml-codex-plugin/clearml/badge)](https://verified-skill.com/skills/sergeinikolenko/clearml-codex-plugin/clearml)

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