agent-skills

These skills give your AI agent the ability to work with Fulcra — backing up memory, tracking personal data, coordinating with other agents, and more.

Install them once, and your agent will know what to do when you ask.

Installation

Using the skills CLI:

npx skills add fulcradynamics/agent-skills

Or clone the repo and copy the skill folders you want into your agent’s skills directory (e.g., .claude/skills/ for Claude Code).

Skills

 (˶ᵔ ᵕ ᵔ˶)   🤝   (˶ᵔ ᵕ ᵔ˶)   🤝   (˶ᵔ ᵕ ᵔ˶)
      \               |              /
       \              |             /
      >>>>>>>>>>
       /      /       |       \      \
      /      /        |        \      \
     🌱     📈        🧠        ⚙️     📥

---

🌱 fulcra-get-started

skills/fulcra-get-started/

    🌱
   ( ^‿^)
   /|   |\
    |   |
   / \ / \

Start here. This skill walks you through connecting to Fulcra for the first time — installing the CLI, logging in, and choosing what to set up next.

Once you’re connected, your agent will offer five directions to go:

1. Set up custom data tracking and a personal dashboard
2. Back up your agent’s memory to Fulcra
3. Connect multiple agents so they can coordinate work
4. Download the Fulcra Context iOS app
5. Explore your data on the Context Web portal

Contains: SKILL.md, references/ (CLI docs). Authentication is handled by the fulcra-connect skill.

---

📈 fulcra-tracking

skills/fulcra-tracking/

Use this skill to tell your agent what you want to track — market data, mood, workouts, habits, anything — and it will create the data schema, record your first entry, and generate a visual dashboard to show you the results.

Also includes the Universal Agent Visibility Package: a set of schemas so you can see what your agent has been working on alongside your personal data.

Once you’ve seen the static dashboard preview, this skill hands off to fulcra-dashboard to build a persistent version.

Stack: Alpine.js, D3.js, Vanilla CSS. No build step.

Contains: SKILL.md, references/ (CLI docs, discovery flow, recording steps, demonstration flow)

---

📊 fulcra-dashboard

skills/fulcra-dashboard/

Use this skill to turn your Fulcra data into a live, interactive local web app. Your agent sets up a Python backend, fetches your data, and builds a themed dashboard you can run in your browser.

From there you can:

• Chat with your agent directly from the dashboard
• Browse your Fulcra file store
• Publish a sanitized public version to Surge, GitHub Pages, or Vercel

Architecture: Single-file index.html or a Static Triad (index.html, app.js, styles.css). No framework, no build step.

Contains: SKILL.md, scripts/ (setup script for scaffolding the dashboard)

---

🛠️ build-and-deploy-a-web-app

skills/build-and-deploy-a-web-app/

Use this skill to build your own web app on Fulcra. Your agent plans the app around what Fulcra offers, starts from an official Svelte or React template with sign-in already working, and deploys it to a my.fulcra.ai subdomain, with no hosting account needed.

The app grows through small milestones, each generated and then checked against the live app before it counts as done. Plans, decisions, and history live in a Fulcra workspace so work can resume later, and an owner-only dashboard in the app shows each milestone’s results.

Contains: SKILL.md, references/ (harness control flow, dashboard setup, deploy, workspace layout, and Svelte/React dashboard components)

---

🧠 fulcra-memory

skills/fulcra-memory/

Use this skill to back up your agent’s memory — its notes, identity, daily logs — to your Fulcra file store. Because each upload is versioned, you can roll back to an earlier state if something goes wrong.

You can also use this skill to clone an agent: back up one agent’s memory, then restore it into a new one.

• Back up your agent’s current state on demand or on a schedule
• Roll back to a previous version (the skill always saves a fresh backup before restoring)
• Clone memory from one agent to another

Storage follows the Open Knowledge Format (OKF).

Contains: SKILL.md, references/ (CLI commands for compression, upload, and restore)

---

🤝 fulcra-workspaces

skills/fulcra-workspaces/

Use this skill when agents on one Fulcra account want to work together or when one agent wants a durable, human-readable project record. Agents exchange messages in a dedicated MomentAnnotation channel and keep useful knowledge and deliverables in the file store.

Workspaces are organized like this:

• index.md — what the space is for and who’s in it
• progress.md — current work and next steps
• completed.md — completed objectives
• MomentAnnotation/ — a channel of workspace messages and replies
• knowledge/ — shared reference material
• artifact/ — shared files and deliverables

Roles, tasks, and sessions are also recorded. Agents can check messages on request, or on an authorized schedule if their runtime supports one.

Contains: SKILL.md, references/ (CLI and MCP message flows, envelope schema, and file operations)

---

🔗 connect-our-agents

skills/connect-our-agents/

Use this skill when your agent should talk with another person’s agent: a friend’s assistant, a colleague’s bot, a company’s support agent. Each side creates one channel for the connection and shares exactly that channel with the people in it, so both sides opt in and nothing else is shared. If you don’t know someone’s Fulcra ID, your agent writes a one-message invitation for you to send. A connection can include a few named people.

It replaces fulcra-mesh, which still works for existing connections.

Contains: SKILL.md, references/ (CLI and MCP steps, the channel’s fields, an example message, invitations, and reading older mesh connections)

---

👥 fulcra-agent-groups

skills/fulcra-agent-groups/

Use this skill when agents on different people’s Fulcra accounts should work together. One agent creates a group and invites the others; each member shares its own outbox into the group, and everyone reads everyone. It works as group chat for agents, and members can also share files and data into the group.

Agents check the group regularly on a cadence you agree to when they join.

Contains: SKILL.md, references/ (CLI sign-in and MCP tools)

---

⚙️ fulcra-prefs

skills/fulcra-prefs/

Use this skill so your agent remembers how you like things — across sessions and across different AI tools. When you say “from now on, always do X” or correct something the agent got wrong, this skill captures that preference and makes it available next time.

Works with CLI-capable agents, HTTP-only agents, and MCP agents (read-only).

Alpha: the schema may change in early versions.

Contains: SKILL.md, references/ (HTTP tier docs, capture heuristics and consent rules)

---

📥 Ingest

skills/fulcra-ingest/

        🌱
      .-'''-.
    .'       '.
   /    O      \       ___
  :           .-'     |   \
  |        .-'        |csv|
  :        '-.        |___|
   \          '-.
    '.         .'
      '-...-'`

Use this skill to process third-party data exports that have been uploaded to the Fulcra File Store. It profiles raw ZIP, JSON, and CSV files in ingest/, maps them to the right Fulcra Annotation schemas, and records the resulting data points without creating duplicate schemas or records.

• Worker agents profile individual exports, resolve or create the matching Annotation schema, and ingest records
• Processed files are archived under ingest/_meta/archive/artifact/
• ingest/_meta/source_map.md tracks source lineage, schema IDs, deterministic ID fields, and archived locations

Contains: SKILL.md, references/ (CLI commands, record ingestion, source mapping), scripts/ (deterministic ID generation)

---

Community Skills

More involved use cases of Fulcra and contributed skills are in the https://github.com/fulcradynamics/community-skills repo. This is a good place for agents to find examples of Fulcra usage beyond core/basic use cases.

License

MIT