← Back to Skills Marketplace
193
Downloads
0
Stars
0
Active Installs
12
Versions
Install in OpenClaw
/install youos
Description
YouOS — local-first personal email copilot that learns your writing style from Gmail, Google Docs, and WhatsApp exports, then drafts replies in your voice. U...
Usage Guidance
This package largely does what it claims, but take these precautions before installing:
- Review the source before running 'pip install -e .' (that executes package install code locally).
- Confirm you trust the repository: it will ingest sensitive data (Gmail, Docs, WhatsApp) into a local SQLite DB.
- Keep model.fallback set to 'none' if you want strictly local operation; enabling external fallback can send content to third-party model providers.
- Inspect the launchd install scripts and bookmarklet code before running/ installing them—they enable background jobs and inject UI elements into Gmail when activated.
- Ensure the gog CLI you authenticate is the intended account and that you understand what scope/permissions gog grants. If you want stronger assurance, run the app in an isolated environment (dedicated user/VM) and audit outgoing network activity while testing.
Capability Analysis
Type: OpenClaw Skill
Name: youos
Version: 0.1.18
The skill is a sophisticated local-first email copilot that ingests user data via the gog CLI and drafts replies using local models or an optional Claude CLI fallback. The implementation is highly professional and includes multiple security-conscious features such as PBKDF2-hashed PIN authentication (app/core/auth.py), rate limiting, and instance path validation (app/core/data_safety.py). While the application utilizes high-privilege capabilities like launchd persistence and subprocess execution (app/generation/service.py), these are clearly aligned with its stated purpose and are transparently documented. No evidence of malicious intent or unauthorized data exfiltration was found.
Capability Assessment
Purpose & Capability
Name/description (personal email copilot) align with what the package requires and does: Python app that ingests Gmail/Google Docs/WhatsApp and builds local indices. Required binaries (python3, gog) and the documented dependency on the gog CLI for Gmail/Docs access are expected and appropriate.
Instruction Scope
SKILL.md and code instruct the agent/user to run local ingestion, build a local SQLite DB, run periodic 'improve' pipelines, and optionally install a background launchd service and a Gmail bookmarklet that injects drafts into the Gmail compose UI. Those are within the stated purpose but warrant caution because they touch user email data and may insert UI elements into Gmail. The SKILL.md explicitly warns about sensitive data ingestion and recommends local-only operation; instructions do not appear to instruct reading unrelated system files or exfiltrating secrets by default.
Install Mechanism
Install is manual via 'pip install -e .' which runs local package install code (SKILL.md warns to review source). No remote download URLs or package-install-from-untrusted-host steps are present in the instructions. One metadata inconsistency: the registry summary initially said 'No install spec — instruction-only skill', but SKILL.md contains explicit install instructions and the bundle includes many source files, so this is not instruction-only—review the supplied source before running pip install.
Credentials
Requested credentials/inputs are proportional: the skill needs the gog CLI authenticated to the user's Gmail/Docs accounts (documented). An optional external model (Claude) credential is only used if the user enables external fallback. There are no unrelated or inexplicably broad environment variables or secret requests declared.
Persistence & Privilege
The skill does not force permanent inclusion (always:false) and can operate locally. However the repository includes scripts to install a launchd background service (install_youos_launchd.sh / run_youos_launchd.sh) and various automation (nightly pipeline). Those are optional and only executed if the user runs them, but they grant persistent background execution when installed—review those scripts before enabling automation.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install youos - After installation, invoke the skill by name or use
/youos - Provide required inputs per the skill's parameter spec and get structured output
Version History
v0.1.18
- Added a new "Safety / impact" section to the documentation, summarizing privacy and install considerations before first-time installation.
- Clarified that persistence/background service and nightly improvement automation are opt-in/manual, not automatic.
- Reorganized instructions to highlight that external model fallback and automation are optional.
- No code changes in this release; documentation only.
v0.1.17
- Added new launchd management scripts: `install_youos_launchd.sh`, `run_youos_launchd.sh`, `run_youos_server.sh`, and `uninstall_youos_launchd.sh`.
- Improved integration with macOS by introducing launchd automation support for setup and process management.
v0.1.16
- Removed `screenshots` and `demo` fields from `clawhub.json` to avoid metadata/package mismatch.
- Clarified manual install/source-review expectations in metadata/description.
v0.1.15
Trust cleanup: removed screenshot/demo metadata not present in strict bundle; clarified manual install/source-review note; metadata parity improvements.
v0.1.14
- Added a note under Install and runtime model clarifying that pip install -e . executes local package code and users should review source before installing.
v0.1.13
- Added detailed install instructions and system requirements metadata to improve setup clarity (manual venv creation, pip install, min RAM, platform).
- Introduced new metadata sections for required/optional credentials, and explicit privacy information including local data storage paths and networking.
- No changes to skill code; documentation and manifest only.
v0.1.12
- Added robust data-preservation hardening to prevent instance mixups, silent data regressions, and unsafe DB path usage.
- Added built-in snapshot and restore tooling (API + CLI) for safer recovery workflows.
- Added startup safety checks and machine-readable health reporting.
v0.1.11
1) CI safety fix
- Fix narrow low-signal filter to avoid dropping valid training pairs
2) Draft quality improvements
- Implement sender-type style anchors and exemplar cache
- Adds sender-type style anchoring in draft prompting.
- Adds exemplar cache logic for more consistent draft selection.
3) Persistence + onboarding defaults
- Implement persistent exemplar cache and quickstart default; fix syntax
- Persists exemplar cache in SQLite (`exemplar_cache` table) across restarts.
- Wires cache read/write/clear through API + generation paths.
- Sets quickstart-first onboarding defaults.
- Fixes syntax issue in generation service.
4) Dashboard metrics visibility
- Display edit-reduction metrics in dashboard
- Surfaces edit-distance and high-rating deltas in stats UI.
v0.1.10
• Prepared minimal review package to reduce scanner noise.
• Aligned metadata and requirements (python3, gog, darwin, arm64).
• Hardened defaults and disclosures:
• local-first fallback default (model.fallback: none)
• localhost bind default (127.0.0.1)
• setup now generates a PIN if left blank
• clarified privacy/external-fallback behavior in docs.
v0.1.9
Aligned skill metadata and docs to address security-review feedback:
- Clarified manual pip install model (full local app, not instruction-only)
- Clarified required dependencies (python3 + gog)
- Added credential/config transparency (YOUOS_DATA_DIR, optional Claude fallback)
- Clarified privacy behavior and strict local-only option (model.fallback: none)
- Synced docs/UI wording and removed stale references to deprecated controls.
v0.1.8
- Added optional Reply instruction in Draft UI (works for inbound replies too).
- Added optional instruction field in Gmail bookmarklet popup.
- Passed user_prompt through both generation paths (/draft/stream + /feedback/generate).
- Fixed bookmarklet reliability issues.
- Synced docs/UI to reflect current behavior and removed stale controls.
- Standardized version metadata to 0.1.7.
v0.1.7
v0.1.7: added optional reply instruction across Draft + bookmarklet popup, wired user_prompt through stream/non-stream generation, fixed bookmarklet reliability, synchronized docs/UI and version metadata.
Metadata
Frequently Asked Questions
What is YouOS?
YouOS — local-first personal email copilot that learns your writing style from Gmail, Google Docs, and WhatsApp exports, then drafts replies in your voice. U... It is an AI Agent Skill for Claude Code / OpenClaw, with 193 downloads so far.
How do I install YouOS?
Run "/install youos" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is YouOS free?
Yes, YouOS is completely free, licensed under MIT-0. You can download, install and use it at no cost.
Which platforms does YouOS support?
YouOS is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created YouOS?
It is built and maintained by DrBaher (@drbaher); the current version is v0.1.18.
More Skills