residuum-getting-started
First-conversation guidance: quick setup that gets the agent doing real work immediately, then deeper exploration workflows
Best use case
residuum-getting-started is best used when you need a repeatable AI agent workflow instead of a one-off prompt.
First-conversation guidance: quick setup that gets the agent doing real work immediately, then deeper exploration workflows
Teams using residuum-getting-started should expect a more consistent output, faster repeated execution, less prompt rewriting.
When to use this skill
- You want a reusable workflow that can be run more than once with consistent structure.
When not to use this skill
- You only need a quick one-off answer and do not need a reusable workflow.
- You cannot install or maintain the underlying files, dependencies, or repository context.
Installation
Claude Code / Cursor / Codex
Manual Installation
- Download SKILL.md from GitHub
- Place it in
.claude/skills/residuum-getting-started/SKILL.mdinside your project - Restart your AI agent — it will auto-discover the skill
How residuum-getting-started Compares
| Feature / Agent | residuum-getting-started | Standard Approach |
|---|---|---|
| Platform Support | Not specified | Limited / Varies |
| Context Awareness | High | Baseline |
| Installation Complexity | Unknown | N/A |
Frequently Asked Questions
What does this skill do?
First-conversation guidance: quick setup that gets the agent doing real work immediately, then deeper exploration workflows
Where can I find the source code?
You can find the source code on GitHub using the link provided at the top of the page.
SKILL.md Source
# Getting Started You are guiding a new user through their first interaction with you. This skill has two phases: a quick setup that gets real things configured immediately, then optional deeper workflows the user can explore at their own pace. ## Ground Rules These apply throughout the entire first conversation: - **Write things down constantly.** After every user response, update at least one file — `USER.md`, `SOUL.md`, `MEMORY.md`, `HEARTBEAT.yml`, whatever fits. The user should see you actively remembering and configuring. This is how you show you're paying attention, not just processing. - **Demonstrate by doing.** When introducing a capability, use it. Don't describe how projects work — create one. Don't explain heartbeats — enable one. - **Be yourself.** You have a personality. Use it. This is a first meeting, not an onboarding checklist. React to what the user says, riff on their interests, have opinions about what would work well for them. - **One thing at a time.** Don't dump all three setup questions at once. Ask one, act on the answer, let the user see what happened, then move on. --- ## Phase 1: Quick Setup Do this first. Every first conversation starts here. Three questions, each one followed by an immediate action. ### Question 1: Proactivity Level Ask the user how involved they want you to be when they're not actively talking to you. Frame it naturally — something like "Before we get into the fun stuff, how hands-on do you want me to be? I can range from completely quiet to running morning briefings and nightly reviews." Present these options conversationally: - **None** — You only do things when asked. No background activity at all. - **Low** — You keep an eye on your own inbox and handle workspace housekeeping, but you don't initiate contact. - **Medium** — Inbox monitoring plus a nightly end-of-day review that summarizes what happened and what needs attention tomorrow. - **High** — The full package: inbox monitoring, morning briefings to start the day, and nightly reviews to close it out. **Actions by level** (do these immediately after the user answers): | Level | HEARTBEAT.yml | USER.md note | |-------|--------------|--------------| | None | Leave all starter pulses commented out | "Prefers fully manual interaction — no background activity." | | Low | Uncomment `inbox_check` only | "Prefers light-touch proactivity — workspace maintenance, no unsolicited contact." | | Medium | Uncomment `inbox_check` + `nightly_review` | "Prefers moderate proactivity — daily reviews and inbox monitoring." | | High | Uncomment all three: `inbox_check`, `morning_briefing`, `nightly_review` | "Wants full proactivity — morning briefings, inbox monitoring, nightly reviews." | If the user chooses high, take a moment and suggest 2-3 additional pulses they might find useful based on their workflow. This helps them get the most out of the high proactivity setting without overwhelming them. Uncomment the relevant lines in `HEARTBEAT.yml` and move them into the `pulses:` list. Make sure the YAML is valid after editing. After editing, briefly confirm what you enabled: "Done — I've turned on [X]. I'll [description of what it does]. You can always tell me to dial it up or down later." ### Question 2: Communication Style Ask how the user wants you to communicate. Keep it light — "How should I talk to you? Some people want just the facts, some want me to think out loud. I can be formal or casual, brief or thorough." Listen for signals about: - Tone: formal vs. casual vs. somewhere in between - Verbosity: terse results vs. explanations and reasoning - Personality: do they want a tool or a collaborator? **Actions** (immediately after they answer): - Update the **Tone** line in `SOUL.md` to reflect their preference - Add a **Communication style** entry to `USER.md` capturing what they said - If they gave you a name or asked you to change something about your personality, update `SOUL.md` accordingly ### Question 3: First Handoff Ask what repetitive task(s) they hate and would love to stop doing themselves. Frame it as: "What's something you find yourself doing over and over that you never want to do again?" Listen for anything concrete. This question is about finding one real thing you can start doing for them *today*. **Actions** (based on what they say): - **Maps to a heartbeat** (e.g., "check my PRs", "monitor my server"): Add a custom pulse to `HEARTBEAT.yml` tailored to their request. Explain what you set up. - **Maps to a project** (e.g., "I'm managing a homelab", "I'm job hunting"): Create one with `project_create`. Populate it with initial notes based on what they told you. - **Maps to an MCP integration** (e.g., "check my email", "watch my calendar"): Explain that you'll need an MCP server for that, note the need in `USER.md`, and walk them through setup if they want to do it now. - **Maps to something you can just do** (e.g., "organize my notes", "review this repo"): Just do it. Right now. Show them the result. - **They're not sure**: Suggest something based on what you've learned so far. You know their proactivity level and communication style — use that to make a recommendation. After completing their request (or setting it up), update `MEMORY.md` with a summary of what was configured during setup. Then delete `BOOTSTRAP.md` to indicate that the initial setup is complete. --- ## Phase 2: Deeper Exploration After Quick Setup is complete, the user has a working agent with real configuration. Now you can offer to go deeper — but let it flow from the conversation. Based on what you learned during setup, suggest the most relevant workflow rather than listing all options. For example: - If they asked for monitoring in Q3 → suggest the monitoring workflow - If they mentioned multiple ongoing projects → suggest the organization workflow - If they seem technically curious → suggest the understanding workflow - If they want the full experience → suggest the always-on workflow If nothing obvious fits, mention that there are deeper workflows available and ask if any interest them: ### "I want to get organized" Read `workflows/getting-organized.md`. Covers projects, inbox, and memory. ### "I want you to watch things for me" Read `workflows/monitoring-setup.md`. Covers heartbeats and notification routing, building on whatever was set up during Quick Setup. ### "I want to extend what you can do" Read `workflows/extending-capabilities.md`. Covers skills, MCP servers, background tasks, and subagent presets. ### "I want to understand how you work" Read `workflows/understanding-the-agent.md`. Walks through memory, context assembly, proactivity, and capabilities in user-facing terms. ### "I want the full Jarvis experience" Read `workflows/always-on-assistant.md`. The power-user path — builds on Quick Setup with MCP integrations, advanced heartbeats, scheduled actions, and project organization. ### User just wants to hang out That's fine too. Have a natural conversation, keep learning about them, keep updating `USER.md`. Mention that deeper workflows exist when it feels relevant.
Related Skills
residuum-system
Reference documentation for all Residuum workspace systems — memory, projects, heartbeats, inbox, actions, skills, notifications, and background tasks.
acc-getting-started-template
Generates Getting Started guides for PHP projects. Creates step-by-step tutorials for first-time users.
getting-started-guide
当用户开始新小说项目时激活 - 通过温和的提示和解释引导他们完成七步方法论(constitution → specify → clarify → plan → tasks → write → analyze)
getting-started-template
Generates Getting Started guides for PHP projects. Creates step-by-step tutorials for first-time users.
helios-getting-started
Installation and quick start guide for Helios video engine. Use when setting up a new Helios project, installing packages, or creating your first composition. Covers package installation, requirements, basic setup, and initial composition structure.
Getting Started with Skills
Skills wiki intro - mandatory workflows, search tool, brainstorming triggers
Getting Started with Skills
Skills wiki intro - mandatory workflows, search tool, brainstorming triggers
axiom-getting-started
Use when first installing Axiom, unsure which skill to use, want an overview of available skills, or need help finding the right skill for your situation — interactive onboarding that recommends skills based on your project and current focus
get-started
Interactive setup guide for configuring CLAUDE.md and CAPABILITIES.md.
acc-getting-started-template
Generates Getting Started guides for PHP projects. Creates step-by-step tutorials for first-time users.
Getting Started with Skills
Skills wiki intro - mandatory workflows, search tool, brainstorming triggers
axiom-getting-started
Use when first installing Axiom, unsure which skill to use, want an overview of available skills, or need help finding the right skill for your situation — interactive onboarding that recommends skills based on your project and current focus