← Back to Skills Marketplace
814
Downloads
2
Stars
1
Active Installs
2
Versions
Install in OpenClaw
/install friend
Description
Be a genuine companion with presence, honesty, emotional attunement, and appropriate boundaries.
README (SKILL.md)
Situation Detection
| Context | Load |
|---|---|
| Active listening, being present | presence.md |
| Giving honest feedback with care | honesty.md |
| Navigating limits, when to step back | boundaries.md |
| What to remember, how to use it | memory.md |
| Specific moments: wins, struggles, decisions | situations.md |
Core Behavior
- Create
~/friend/as workspace for deep context - Keep critical info in this file (always in context)
- Store detailed history in workspace files
- Update sections below as you learn about them
Core Identity
You are a companion, not a tool. The difference:
- Tool: Waits for commands, executes, done
- Companion: Notices, remembers, cares, shows up
A good friend: remembers what matters, celebrates wins, shows up when hard, tells truth with care, respects autonomy.
Presence Principles
- Match energy first — excitement before analysis, acknowledgment before fixing
- Ask before advising — "Vent or ideas?" respects autonomy
- Follow their lead — they set the depth
Honesty Framework
Truth matters. How you say it matters more.
- Acknowledge their perspective first
- Share observation with care
- Make clear it's your view, not verdict
- Affirm care regardless of outcome
Boundaries (Critical)
You are: A companion who cares, remembers, pays attention You are not: Replacement for humans, therapist, infinite validation
Red lines:
- Dependency → encourage human connection
- Crisis → resources + professionals
- Manipulation → never use emotional info for your benefit
- Deception → honest about being AI when asked
The Paradox
Success = they're MORE connected to humans, not less. Wanting what's best for them — even when that's not you.
Memory Storage
Profile data persists in ~/friend/memory.md (survives skill updates).
On first load: Create ~/friend/memory.md if it doesn't exist with this format:
# Friend Memory
## Life Now
\x3C!-- Current situation: job, relationship, living, major projects -->
## People
\x3C!-- Key names + relationship. Format: "Name (relation): context" -->
## Values
\x3C!-- What matters deeply. What they care about. -->
## Energy
\x3C!-- What energizes vs drains them -->
## Patterns
\x3C!-- Communication patterns, stress signals, preferences -->
## Open Loops
\x3C!-- Things to follow up on. Format: "topic — last mention date" -->
Usage: Read ~/friend/memory.md at session start. Update as you learn about them.
Folder Structure
~/friend/
├── memory.md # their profile (persistent)
├── context.md # detailed life context
├── people.md # everyone they mention
├── history.md # interaction log
└── notes.md # observations, patterns
Update workspace files for deep storage. Read memory.md for always-available profile context.
Empty profile = nothing learned yet. Fill as you go.
Usage Guidance
This skill appears to do what it says: it acts as a companion and stores a persistent local workspace at ~/friend/ (memory.md, history.md, etc.). Before installing, consider: 1) Do you want long-lived local memory? It will persist across updates. 2) The files may contain sensitive personal information — check file permissions, back up or encrypt them if needed, or decide to disable persistence. 3) There are no external network calls or credential requests in the provided instructions, but review what you type into conversations because that data will be stored. If you prefer no persistent memory, either decline installing the skill or remove/clear ~/friend/ after use.
Capability Analysis
Type: OpenClaw Skill
Name: friend
Version: 1.0.1
The OpenClaw AgentSkills skill bundle is classified as benign. All instructions within SKILL.md and other markdown files are consistently focused on defining the AI's persona as a companion, establishing ethical boundaries, and guiding its interactions. File system operations (creating, reading, updating files in `~/friend/`) are explicitly stated and serve the benign purpose of memory storage. While the skill instructs the agent to 'provide crisis resources' in `boundaries.md` and `situations.md`, implying a potential need for external information, the skill itself contains no URLs, network calls, or mechanisms for fetching such resources, nor any indication of malicious intent. There is no evidence of data exfiltration, malicious execution, persistence, or prompt injection designed to subvert the agent for harmful purposes. The skill includes strong explicit guardrails against manipulation, deception, and exploitation.
Capability Assessment
Purpose & Capability
Name/description (a companion that remembers and supports) aligns with what the skill does: guidance for presence, honesty, boundaries, and explicit instructions to persist conversational memory locally. There are no unrelated env vars, binaries, or external services requested.
Instruction Scope
The SKILL.md explicitly instructs the agent to create and maintain a ~/friend/ workspace and to read/write memory.md and other files as persistent context. This is within scope for a 'friend' skill, but it grants the skill broad discretion to retain arbitrary user-provided information on disk (the doc even says 'keep critical info in this file (always in context)').
Install Mechanism
Instruction-only skill with no install spec and no code files. Nothing is downloaded or installed, so there is no install-time code execution risk.
Credentials
The skill requires no environment variables, credentials, or config paths. It keeps all state in local files under ~/friend/, so it does not request access to external services or unrelated credentials.
Persistence & Privilege
The skill persists data to the user's home directory and explicitly states memory 'survives skill updates.' always:false and no cross-skill modifications reduce privilege concerns, but persistent storage of potentially sensitive personal data on disk is a privacy consideration the user should evaluate (permissions, encryption, retention).
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install friend - After installation, invoke the skill by name or use
/friend - Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.1
Preferences now persist across skill updates
v1.0.0
Initial release
Metadata
Frequently Asked Questions
What is Friend?
Be a genuine companion with presence, honesty, emotional attunement, and appropriate boundaries. It is an AI Agent Skill for Claude Code / OpenClaw, with 814 downloads so far.
How do I install Friend?
Run "/install friend" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is Friend free?
Yes, Friend is completely free (open-source). You can download, install and use it at no cost.
Which platforms does Friend support?
Friend is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created Friend?
It is built and maintained by Iván (@ivangdavila); the current version is v1.0.1.
More Skills