AgentToolDrop

19 of 57

SkillOpt

Trains a skill document for a frozen LLM agent with validation-gated edits, shipped as best_skill.md

  • score 48
  • 18.5k stars
  • 114 stars/day since creation
  • pushed
  • v0.2.0 on
  • MIT

Install from PyPI

python -m pip install skillopt

Copied from the project's docs at github.com/microsoft/SkillOpt/blob/main/docs/guide/installation.md

Why you'd want it

SkillOpt treats the skill document as the trainable state of a frozen agent and trains it without touching model weights. A separate optimizer model turns scored rollouts into bounded add, delete or replace edits, and in the default path an edit is accepted only when it strictly improves a held-out validation score. The deployed artifact is a best_skill.md file that runs against the unchanged target model, and the README says it adds zero inference-time model calls at deployment. A skillopt-sleep CLI, labelled preview, applies the same gate to past coding-agent sessions and stages proposed updates for you to adopt.

Best for

  • Training a skill file against scored tasks
  • Tuning skills for Codex or Claude Code runs
  • Reviewing past coding-agent sessions into skills (preview)

Use it when

You have tasks you can score and a held-out split, and you want a skill document for an agent whose model you keep frozen improved by measured edits instead of hand-tuning.

Skip it if

You have no scored tasks or validation split to gate edits against, or no access to a configured model backend for the rollout and optimizer calls. The README calls SkillOpt-Sleep a preview and some multi-objective, replay and dream-rollout controls experimental, and the docs say its interfaces and defaults may change. With a real backend, Sleep sends truncated excerpts of harvested session transcripts to the provider you pick, and the docs say outbound prompts are not guaranteed secret-free, so review transcripts first.

Editor's note

PyPI serves 0.2.0. The docs say some newer features, such as Cursor and Pi support and the openai_compatible research backend, need a source install from main, and the benchmark configs and agent integration files live in the repository rather than the wheel.

Related tools

Same category or shared tags.

  1. Rank 2

    OpenFilm

    Devtool

    Lets a coding agent make videos as HTML pages, with a local browser editor to change anything it makes

    • score 76
    • 909 stars
    • 64.5 stars/day since creation
    • pushed
    • v0.2.0 on
    • MIT

    Add the OpenFilm skill to your agent

    npx skills add openfilm/openfilm
    Use it when, skip it if, stats for OpenFilm

    Why you'd want it

    OpenFilm has a coding agent such as Codex or Claude Code write a film as a folder with a film.html that cuts animated web pages, footage, pictures and sound together on tracks, styled with CSS. Each shot is a web page that draws any moment of itself with frame(t), and the same t must always draw the same picture. Studio, a local editor that runs in your browser, lets you move, scale and rotate elements, change text and color, and trim clips on the timeline, and the edits are saved to the film's files so the agent sees and keeps them. The look command checks the film by drawing each frame of its sheet twice and naming any that changed, along with page errors, failed fonts and missing sound files, and render writes the MP4.

    Best for

    • Agent-made launch videos with music
    • Hand-editing an agent's film in Studio
    • Exporting MP4, GIF or timelines for Premiere, Resolve and Final Cut

    Use it when

    You already work with Claude Code, Codex or another coding agent and want it to produce a video that you can then adjust by hand in an editor.

    Skip it if

    You want a polished, long-stable tool: the repo is about two weeks old and its releases are 0.x (latest v0.2.0). Setup needs Node 22.2+, ffmpeg and git, and the first run downloads about 100 MB of headless Chromium, though the desktop app bundles ffmpeg and the browser. Your agent runs openfilm commands and the pages render in a headless browser, and generated voice, music, images and video need your own provider keys entered in Studio, so review what the agent runs and where keys go.

    Stats

    New and rising score
    76 of 100
    Stars per day, since repo creation
    64.5
    Release
    v0.2.0,
    License
    MIT
    Pricing
    Open source
    Works with
    Claude Code, Codex, macOS, Linux, Windows
    Forks
    95
    Open issues
    7
    GitHub checked

    Full stat block for OpenFilm

  2. Rank 5

    text-to-cad

    Devtool

    Gives an agent skills to generate 3D models as STEP, STL, 3MF or GLB and run manufacturing checks

    • score 62
    • 21.4k stars
    • 268 stars/day last 3 days
    • pushed
    • v0.7.22 on
    • MIT

    Install the Claude Code plugin

    claude plugin marketplace add earthtojake/text-to-cad#latest
claude plugin install text-to-cad@earthtojake
    Use it when, skip it if, stats for text-to-cad

    Why you'd want it

    text-to-cad is a plugin and skill set that gives an agent local workflows for generating 3D models as STEP, GLB, STL or 3MF files. Its skills also cover design for manufacturing checks, engineering drawings, 2D DXF files, URDF, SRDF and SDF robot and simulator files, and slicing to G-code with OrcaSlicer. Models open in a CAD Viewer, shown as a card in the conversation where the app supports it or as a browser link otherwise. The README says it works with agents that support plugins or the skills framework, including Claude Code, Codex, Cursor, Gemini and Grok.

    Best for

    • Prompting an agent to model printable parts
    • DFM and printability checks before fabrication
    • Robot description files for URDF and SDF

    Use it when

    You want a coding agent to produce and edit CAD files from plain-language or image requests, then check them for 3D printing, sheet metal or CNC before sending them out.

    Skip it if

    You do not want telemetry: usage stats and crash reports are on by default with a random install ID, and you turn them off with uvx cadgen telemetry off. It needs uv installed, downloads its CAD runtime on first start, and runs a local server and a browser viewer. Its skills can send prints to Bambu Lab printers, so review what an agent may do there. Releases are still 0.x (v0.7.21), and on Windows 11 with Smart App Control on, the README says CAD commands fail unless you turn it off or use WSL.

    Stats

    New and rising score
    62 of 100
    Stars per day, last 3 days
    268
    Release
    v0.7.22,
    License
    MIT
    Pricing
    Open source
    Works with
    Claude Code, Codex, Cursor, Gemini CLI, Claude Desktop
    Forks
    2,108
    Open issues
    47
    GitHub checked

    Full stat block for text-to-cad

  3. Rank 7

    SkillSpector

    Devtool

    Scans agent skills for prompt injection, data exfiltration and supply-chain risks before you install them

    • score 60
    • 20.5k stars
    • 71.9 stars/day last 3 days
    • pushed
    • v2.13.0 on
    • Apache-2.0

    Install with uv

    uv tool install git+https://github.com/NVIDIA/skillspector.git
    Use it when, skip it if, stats for SkillSpector

    Why you'd want it

    SkillSpector scans a skill from a Git repo, URL, zip file, directory or single SKILL.md against 71 patterns in 17 categories, including prompt injection, data exfiltration, privilege escalation, MCP tool poisoning and MCP least privilege. It runs static analysis first and an optional LLM evaluation second, gives a 0 to 100 risk score, and writes terminal, JSON, Markdown or SARIF reports. It can also run as an MCP server with a scan_skill tool, so Claude Code, Codex CLI or Gemini CLI can gate skill and MCP installs on the result. It is part of NVIDIA's Verified Skills pipeline.

    Best for

    • Vetting third-party skills before install
    • CI checks with SARIF output
    • Gating installs from inside an agent

    Use it when

    You install skills or MCP servers from public repositories and want a report on risky patterns before they reach your agent.

    Skip it if

    You only use skills you wrote yourself. The optional LLM analysis sends skill content to the configured provider with your API key; --no-llm runs static checks only. The README notes the stdio MCP transport still has an initialize hang reported in issue #199.

    Stats

    New and rising score
    60 of 100
    Stars per day, last 3 days
    71.9
    Release
    v2.13.0,
    License
    Apache-2.0
    Pricing
    Open source
    Works with
    Claude Code, Codex, Gemini CLI, Docker
    Forks
    1,802
    Open issues
    153
    GitHub checked

    Full stat block for SkillSpector

For agents

The same entry, as files an agent can read without scraping this page. 

# read one entry
curl -s https://agenttooldrop.com/tools/skillopt.md

# list the MCP servers in the catalog
curl -s https://agenttooldrop.com/tools.json \
  | jq -r '.tools[] | select(.category=="mcp-server") | .slug'