Search skills/

intellectronica agent skills Claude skills

A personal collection of 44 mostly small skills: API wrappers for Notion, Todoist, Anki and Context7, image generation, Mermaid rendering and git setup.

intellectronica/agent-skills 44 skills Productivity & notesSkill & agent tooling
Our verdict Recommended

Clean, CC0-licensed and easy to pick from; install only the skills for services you actually use.

2/5 Depth expertise shipped Note Security Review before installing 6mo Maintenance Updated 6 mo ago 96K Adoption installs OK License CC0-1.0, commercial OK

What intellectronica agent skills does

This is the personal skill set of the GitHub user intellectronica, installable one skill at a time through the skills CLI or as a Claude Code plugin marketplace. Most skills wrap a tool or web service so the agent uses it correctly. Examples include anki-connect (talks to a local Anki through AnkiConnect with a rule to ask before changing notes or cards), notion-api, todoist-api (through the td CLI), raindrop-api, tavily, context7 for current library docs, and gog-cli for Google Workspace from the command line. Others generate images with OpenAI's GPT Image 1.5 or Google image models, render Mermaid diagrams to SVG and PNG, convert documents to Markdown, make code snippet images through ray.so, create lorem ipsum, and initialize a git repository.

Skills average about 6.7 KB. A few have reference folders (the Copilot SDK skill has several) and the image and Mermaid skills include Python or TypeScript scripts. There are no evals. Because each skill targets a specific service, the practical value depends on whether you use that service, and several need API keys supplied through environment variables such as OPENAI_API_KEY. The README advertises tutorials and a workshop from the author's company, but the skills do not depend on them.

Good for

  • Users of Notion, Todoist, Anki, Raindrop or Context7 who want the agent to call them correctly
  • Anyone wanting working examples of API-wrapper skills
  • People who like picking single skills instead of installing a large pack

Not for

  • Users who want a coherent workflow or methodology; this is a grab bag
  • Anyone unwilling to give an agent API keys for third-party services

What's inside depth 2/5

44 skills averaging about 6.7 KB with a few reference folders and a handful of scripts, but no tests or evals.

Skills
44
Reference docs
44
Scripts
8
Tests / evals
0
Templates
0
Avg SKILL.md
6.6 KB

Counted from the repository file tree. How we score depth.

Dependencies and costs

  • API keys or accounts for the chosen skills (Notion, Todoist, Tavily, OpenAI, Google, Context7, Raindrop)
  • Anki with the AnkiConnect add-on for anki-connect
  • Node or Python for the scripted skills

Security review

Review before installing

A set of small API-integration and utility skills (Notion, Raindrop, Todoist, Tavily, Context7, Upstash Redis, OpenAI and Gemini image generation, YouTube transcripts, Google Workspace via the gog CLI, Anki, Mermaid rendering, mgrep code search). Each skill sends only its own service's key to that service, and the gog skill tells the agent not to send mail, share publicly or delete without an explicit request. Several skills install packages or run unpinned tools on demand, and the mgrep skill tells the agent to index the repository with a tool whose storage is not documented here.

Worth knowing

  • Needs an API key for its own serviceAPI skills use their own service's key only against that service: UPSTASH_REDIS_REST_TOKEN, OPENAI_API_KEY, GEMINI_API_KEY, TAVILY_API_KEY, and Notion, Raindrop, Todoist and Monologue tokens.skills/upstash-redis-kv/scripts/upstash-client.ts:193
  • Makes network requestsSkills call external services: context7.com (library and query text), Tavily, Notion, Raindrop, Todoist, Monologue, OpenAI and Gemini image APIs, YouTube transcripts, and local AnkiConnect on 127.0.0.1:8765.skills/context7/SKILL.md:19
  • Installs packages or toolsrender.ts runs bun add or npm install for a missing package at run time, and youtube-transcript, markdown-converter and image scripts pull Python packages through uv / uvx.skills/beautiful-mermaid/scripts/render.ts:153
  • Runs remote code at an unpinned versionTells the agent to run bunx @mixedbread/mgrep watch (unpinned) which indexes the repository, and uses uvx markitdown and uv run scripts with youtube-transcript-api>=1.0.0 without exact pins.skills/mgrep-code-search/SKILL.md:33
  • Relies on a CLI or package we could not reviewSeveral skills rely on CLIs and packages not in the repo: gog (Google Workspace, installed with brew install gogcli), mgrep, markitdown, agent-browser, bun/tsx and the Copilot SDK.skills/gog-cli/SKILL.md:28

What we checked

The repository (156 files) on Oct 29, 2026. Not a full review: it relies on code outside the repository: bunx @mixedbread/mgrep (the skill indexes the repository; where the index is stored is not described in the skill); gog / gogcli (Google Workspace CLI, installed from Homebrew); uvx markitdown and Python packages pulled by uv scripts (youtube-transcript-api, openai, google-genai, pillow).

NVIDIA SkillSpector static scan: 424 pattern hits (Data Exfiltration 244, MCP Rug Pull 102, Privilege Escalation 38, Excessive Agency 27, Memory Poisoning 4, Tool Misuse 4). Most hits are API-key and curl examples in skill docs, with Markdown 'Concealed Executable Artifact' notes on the two .github maintainer scripts that regenerate plugin manifests and the README. The plugin and skill folders are duplicated, so each script appears twice; the copies are byte-identical.

A review, not a certification. Skills change after we check them — read the files before installing.

License

CC0-1.0 — public domain dedication, commercial use allowed.

Can it run unattended?

Not really. These are on-demand tool wrappers, several of them told to confirm before modifying data.

Alternatives to intellectronica agent skills

Matt Pocock's Skills

mattpocock/skills

Small, composable engineering skills: grill-me interviews, shared glossary, TDD, triage, prototyping, code review, handoff and a flow router.

Recommended Check first Depth 2/5

Awesome Claude Skills (Composio)

ComposioHQ/awesome-claude-skills

A curated list of Claude skills plus hundreds of templated app-automation skills that run through Composio's hosted Rube MCP.

With caveats Broad access Depth 1/5

Matteo Collina's Skills

mcollina/skills

Eleven skills from a Node.js core maintainer: Fastify, Node internals, TypeScript types, OAuth, docs structure, AGENTS.md and gh workflows.

Recommended Check first Depth 3/5

Claude Code Infrastructure Showcase

diet103/claude-code-infrastructure-showcase

A reference setup for Claude Code: hooks that auto-activate skills, a skill-rules.json config, and Node/Express/React stack guides to copy into your project.

With caveats Broad access Depth 3/5

FAQ

What is intellectronica/agent-skills?

It is a personal collection of 44 agent skills, mostly wrappers for tools and services such as Notion, Todoist, Anki, Context7 and image generation APIs. You install the individual skills you want.

How do I install intellectronica agent skills?

Use npx skills add intellectronica/agent-skills --skill <skill-name>. In Claude Code you can also run claude plugin marketplace add intellectronica/agent-skills and then claude plugin install <skill-name>@intellectronica-skills.

Are these skills safe?

The scripts we read only call the services they advertise and take API keys from your environment or arguments. Skills that modify data, like the Anki one, ask for confirmation first, but each service wrapper does send your content to that service.

intellectronica skills vs mattpocock skills: which is better?

They solve different problems. This repo wraps external services and utilities, while Matt Pocock's skills are about engineering habits and workflow.

Sources: GitHub repository (stars, license, file tree, last push Apr 25, 2026); install counts from skills.sh via Best Skills (CC BY 4.0). Review based on the README and SKILL.md files as published.