claudecodex

probe-research

Experiment-tracking skills (probe-research:probe, probe-research:track-work, probe-research:visualize-progress, probe-research:instrument-code, probe-research:plan-research-review) wired to a read-only Probe MCP server; writes via the `probe` CLI.

0 on GitHubby Probev0.287.0Apache-2.0updated todayDeveloper Tools
/

Plugin probe-research from marketplace research-os-agent in prbe-ai/research-os-agent. Commands read from the repo's manifest — today.

This plugin runs code on your machine — its hooks fire automatically on events like tool use. The exact commands are listed below; read them before installing.

What it installs

Skills (17)

  • artifact-design

    Design a self-contained HTML page as the design lead at a small studio would - palette, type, layout and copy pitched to the subject, never templated. Load before writing any page.

  • artifact-diagramming

    Draw inline SVG diagrams that show a mechanism (what flows where, what differs between options) as the engineer who has to live with the decision. Load before drawing any diagram.

  • audit-team-note

    Periodic team note audit - should be used in conjunction with edit-notes skill

  • edit-notes

    How to write notes on a project, experiment or run, and the team note.

  • instrument-code

    Wires the Probe SDK (`probe.init`, `probe.log`) into a script that trains, evaluates, sweeps or serves, so that it records run-level metrics, and an RL trainer (Miles, ART, SkyRL, verl, TRL) or a Harbor job, so that it records every rollout as a sample. Use when writing or changing any such script,

  • notes-audit

    Renamed to `audit-team-note` - load that skill. This stub ships for one release so a CLI whose audit dispatch still says `notes-audit` resolves to something.

  • overview-charts

    Which charts a page draws and how to frame each on the point it makes. Load when an entity has metrics.

  • overview-contract

    The page's HTML, CSS and script rules; push refuses a page that breaks one. Load before editing a page.

  • overview-page

    How a page is built - its frame, headings, length, parts and a person's blocks. Load before editing a page.

  • overview-visuals

    Pick an experiment's tables and charts, or a project's tables; you name fields and runs, the server reads the numbers. Read before `probe visuals set`.

  • plan-research-review

    The research plan review - maps a research plan to the code that already exists and asks the researcher about every gap; run it only when the researcher asks for it or your Probe instructions say to.

  • probe

    The skill for setting the Probe plugin state - `on` (r/w), `read`, `off` (no r/w).

  • track-work

    Record ML work in Probe - track ALL related work (training runs, evaluation sweeps, evals, prompt engineering, datasets, lit reviews, architecture design work, etc) - not product code that only calls a model, its CI or deploys, or work on Probe's own code - this skill tells HOW to track this data pr

  • visualize-progress

    Renders progress of the current project/work tracked in Probe - meant to be used to help the user visualize what is going on.

  • write-pane

    Write a pane - a live visual on a project or experiment page. Use it unasked whenever a run, experiment or project has important data the built-in charts and tables can't show; never for what they can. When you report such results, put them in a pane too, without asking, since the researcher wants t

  • write-preview

    Write a project's preview - the pane in its box on the canvas that makes plain what the project is about. Use it unasked right after a project is created, and again only when what the project is about changes or something explains it better.

  • write-summary

    Keep each project's and experiment's page - what people read to catch up.

Slash commands (1)

  • /probe-research:probe-research-setup

    Set up Probe on this computer from this chat - sign in, connect the coding agents, and bring in work the team already has (past chats, project folders, W&B runs). Agent-run, idempotent; the person only answers questions and approves in the browser.

MCP servers (1)

  • probe-research

    connects to: https://mcp.research.prbe.ai/mcp

Hooks (7)

  • UserPromptSubmit

    /bin/bash -c 'ROOT="${PLUGIN_ROOT:-${CLAUDE_PLUGIN_ROOT:-}}"; ROOT="${ROOT%/}"; if [ ! -x "$ROOT/hooks/dispatch.sh" ] && [ -n "${PLUGIN_ROOT:-}" ]; then ROOT="$(ls -dt "${ROOT%/*}"/[0-9]*.[0-9]*/.code

  • SessionStart

    /bin/bash -c 'ROOT="${PLUGIN_ROOT:-${CLAUDE_PLUGIN_ROOT:-}}"; ROOT="${ROOT%/}"; if [ ! -x "$ROOT/hooks/dispatch.sh" ] && [ -n "${PLUGIN_ROOT:-}" ]; then ROOT="$(ls -dt "${ROOT%/*}"/[0-9]*.[0-9]*/.code

  • PreCompact

    /bin/bash -c 'ROOT="${PLUGIN_ROOT:-${CLAUDE_PLUGIN_ROOT:-}}"; ROOT="${ROOT%/}"; if [ ! -x "$ROOT/hooks/dispatch.sh" ] && [ -n "${PLUGIN_ROOT:-}" ]; then ROOT="$(ls -dt "${ROOT%/*}"/[0-9]*.[0-9]*/.code

  • PreToolUse

    /bin/bash -c 'ROOT="${PLUGIN_ROOT:-${CLAUDE_PLUGIN_ROOT:-}}"; ROOT="${ROOT%/}"; if [ ! -x "$ROOT/hooks/dispatch.sh" ] && [ -n "${PLUGIN_ROOT:-}" ]; then ROOT="$(ls -dt "${ROOT%/*}"/[0-9]*.[0-9]*/.code

  • PostToolUse

    /bin/bash -c 'ROOT="${PLUGIN_ROOT:-${CLAUDE_PLUGIN_ROOT:-}}"; ROOT="${ROOT%/}"; if [ ! -x "$ROOT/hooks/dispatch.sh" ] && [ -n "${PLUGIN_ROOT:-}" ]; then ROOT="$(ls -dt "${ROOT%/*}"/[0-9]*.[0-9]*/.code

  • SessionEnd

    /bin/bash -c 'ROOT="${PLUGIN_ROOT:-${CLAUDE_PLUGIN_ROOT:-}}"; ROOT="${ROOT%/}"; if [ ! -x "$ROOT/hooks/dispatch.sh" ] && [ -n "${PLUGIN_ROOT:-}" ]; then ROOT="$(ls -dt "${ROOT%/*}"/[0-9]*.[0-9]*/.code

  • Stop

    /bin/bash -c 'ROOT="${PLUGIN_ROOT:-${CLAUDE_PLUGIN_ROOT:-}}"; ROOT="${ROOT%/}"; if [ ! -x "$ROOT/hooks/dispatch.sh" ] && [ -n "${PLUGIN_ROOT:-}" ]; then ROOT="$(ls -dt "${ROOT%/*}"/[0-9]*.[0-9]*/.code

View source on GitHub useprobe.ai

More from prbe-ai/research-os-agent

All 4 plugins →

The daemon profile (installed by `probe wizard` in place of probe-research): the Probe daemon records and reads for your coding agent. Carries instrument-code, plan-research-review, visualize-progress, the daemon's messages and questions, and the guard; no MCP.

16 skills · hooks

0

probe-hooks

claudecodex

Probe's hook proxy, installed by `probe wizard`: hands each hook event to the Probe service on this machine. Its files never change.

hooks

0

Ships client-sanitized per-session Claude Code or Codex transcripts to Probe with source-bound device authentication.

hooks

0

Similar plugins

superpowers

claudecodex

Core skills library for Claude Code: TDD, debugging, collaboration patterns, and proven techniques

15 skills · hooks

obra/superpowers 316,836

Scaffold, validate, and publish new Claude Code plugins with proper structure

3 skills · 1 command · 1 agent

ruvnet/ruflo 81,589

context7

claudecodex

Up-to-date documentation lookup. Pull version-specific documentation and code examples directly from source repositories into your coding agent's context.

1 MCP server

upstash/context7 64,709

Use Alibaba Open Code Review from local Codex.

2 skills

alibaba/open-code-review 50,068

Query Trackio projects, runs, and metrics using the `trackio` CLI. Use when the user needs to list projects/runs/metrics, get summaries, or retrieve metric values from local Trackio databases. Supports both human-readable and JSON output formats. Covers project discovery, run inspection, metric querying, and system metrics. Designed for LLM agents and automation scripts.

1 skill

sickn33/agentic-awesome-skills 49,620

slidev

claude

Claude Code skill for creating and presenting Slidev slidedecks

1 skill

slidevjs/slidev 49,452

Maintain probe-research?

This page updates itself from your repo. Add the badge to your README, or get it Featured to pin it above the organic results with a followed link to your site.

claude & codex plugins4 plugins