โ Back to Skills Marketplace
Proactive Claw
by
Dawid Piaskowski
ยท GitHub โ
ยท v1.2.41
1387
Downloads
3
Stars
13
Active Installs
50
Versions
Install in OpenClaw
/install proactive-claw
Description
๐ฆ The most powerful proactive engine for OpenClaw. Your personal assistant that learns from you and helps you be more productive.
Usage Guidance
This skill appears to do what it says: a local-first proactive calendar assistant that reads your calendars and writes to a dedicated Actions calendar. Before installing, review these items: 1) credentials.json is required and contains your OAuth client secret โ only provide a client you control and understand; token.json will store OAuth tokens locally, revoke them via your Google account if you stop using the skill. 2) The OAuth scope is full calendar access; the code claims to only write to an Actions calendar, but that restriction is enforced in code (so only use trusted copies of the skill). 3) setup.sh requires you to install Python packages manually from requirements-*.txt; consider installing into a virtualenv and prefer hash-locked dependencies. 4) Try the dry-run commands first (scripts/setup.sh --doctor, python3 scripts/action_planner.py --plan --dry-run, python3 scripts/action_executor.py --execute --dry-run, python3 scripts/quickstart.sh) to inspect planned actions before enabling the daemon. 5) If you care about maximum isolation, run the skill in a separate user account or VM, and audit config.json defaults (max_autonomy_level defaults to confirm). If you want additional assurance, request the full contents of the omitted files for manual review or run the audit commands listed in SECURITY.md.
Capability Analysis
Type: OpenClaw Skill
Name: proactive-claw
Version: 1.2.41
Proactive Claw is a comprehensive calendar management engine that demonstrates high-quality security engineering and transparency. Key safety features include a strict write-guard in `scripts/cal_backend.py` that restricts modifications to a specific 'Actions' calendar, local-first data persistence using SQLite (`memory.db` and `proactive_links.db`), and a setup process (`scripts/setup.sh`) that avoids risky patterns like `curl|bash` or automatic dependency installation. The bundle includes extensive documentation (`SECURITY.md`) detailing its threat model and data handling, and the code logic for autonomous actions is governed by user-configurable autonomy levels that default to 'confirm' mode.
Capability Assessment
Purpose & Capability
Name/description (proactive calendar assistant) match the code and runtime requirements: python3, a local config.json, and credentials.json for Google OAuth. The codebase implements calendar scanning, planning, and local notification/execution logic consistent with the stated purpose.
Instruction Scope
SKILL.md instructs running scripts/setup.sh, quickstart.sh and local python scripts (daemon, planner, executor). The runtime flow legitimately reads/writes files under ~/.openclaw/workspace/skills/proactive-claw, runs a local daemon, and performs OAuth calls to Google and optional Nextcloud CalDAV only when configured. Note: write access to a dedicated 'Actions' calendar is implemented in code (and the OAuth scope is full calendar access); the claim that writes target only the Actions calendar is a code-level enforcement, not an OAuth-scope restriction.
Install Mechanism
This is instruction-only (no automated install spec). setup.sh prints manual install commands and requires the user to run pip installs using provided requirements files. The requirements files are standard but do not include hash-locked packages (supply-chain risk if you blindly pip install). No downloads from unknown URLs or extract/install steps in the published core bundle were observed.
Credentials
No environment variables are requested. The skill requires credentials.json (Google OAuth client_id/secret) and writes token.json; this is proportionate for Google Calendar integration but is sensitive โ credentials.json and token.json contain secrets and should be managed carefully. No unrelated credentials or system-wide secrets are requested.
Persistence & Privilege
The skill does not request always:true and defaults to interactive/confirm autonomy. Persistent state is confined to the skill workspace (~/.openclaw/workspace/skills/proactive-claw) and local SQLite files; no system services, sudo, or cross-skill configuration writes were observed.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install proactive-claw - After installation, invoke the skill by name or use
/proactive-claw - Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.2.41
- Improved setup flow with new `setup.sh --doctor` and install command helpers.
- Added `RELEASE_CHECKLIST.md` and requirements files for Google and Nextcloud integrations.
- Introduced `scripts/quickstart.sh` for faster initial preview runs.
- Enhanced documentation for manual dependency installation and backend bootstrapping.
v1.2.40
- Bumped version to 1.2.40.
- Documentation updates: SKILL.md revised (minor formatting, content, or metadata tweaks).
- No functional code changes included.
v1.2.39
- Improved setup instructions: `setup.sh` now validates dependencies before configuration.
- Removed deprecated scripts: `llm_rater.py` and `voice_bridge.py` have been deleted.
- Updated documentation in `SKILL.md`, `CHANGELOG.md`, and `SECURITY.md` for clarity and accuracy.
- Configuration examples and scripts (`config.example.json`, `adaptive_notifications.py`, `config_wizard.py`, `orchestrator.py`) were updated to reflect recent changes and streamline onboarding.
v1.2.37
Core split release: integration-heavy helpers moved to proactive-claw-integrations; core now excludes cross-skill, team-awareness, daemon installer, and clawhub OAuth bootstrap scripts.
v1.2.36
Clean-scan privacy hardening: local-only llm_rater, team_awareness excluded from bundle, docs/config updated.
v1.2.35
Security hardening: optional clawhub OAuth bootstrap now fail-closed with explicit opt-in + required SHA-256 pin verification and strict payload validation.
v1.2.34
Security hardening for voice routing: in-process allowlist dispatch, stricter capture validation, policy intent direct handling.
v1.2.33
- Added new configuration options for backend and mode selection in `config.example.json`
- Improved calendar backend handling in `scripts/cal_backend.py`
- Enhanced setup guidance in documentation
- Updated internal logic for background daemon and planning in `scripts/daemon.py` and `scripts/action_planner.py`
- Improved configuration wizard in `scripts/config_wizard.py`
- Documentation and metadata updates (SKILL.md, CHANGELOG.md)
v1.2.32
**Summary: Refined safety, onboarding, and clarity in documentation.**
- Updated `SKILL.md` with clearer, friendlier descriptions and safety assurances.
- Added quick start instructions and setup presets to improve first-time user experience.
- Provided concise summaries of what users experience in the first week.
- Presented new architecture diagrams and example scenarios in a simpler, more approachable format.
- Maintained all technical and install requirements from previous versions.
v1.2.31
proactive-claw 1.2.31
- Updated SKILL.md flow diagram to clarify the two-way interactions and data flow between calendar and chat.
- Made minor text tweaks in SKILL.md for clarity.
- No changes to core logic or interfaces.
v1.2.30
**Proactive Claw v1.2.30 โ More powerful proactive engine, major documentation and feature updates**
- Expanded and clarified technical and safety documentation: install, config, local-only guarantees, architecture overview.
- Significantly updated SKILL.md for clearer user guidance, scenarios, and FAQ.
- Added new core scripts for cross-skill interaction, LLM-based scoring, team awareness, and voice bridging.
- Improved proactive loop: now includes built-in lightweight chat scoring and smarter suggestions.
- Installation and background execution process is now simpler and more explicit.
- All state and credentials remain local by default; autonomy stays opt-in.
v1.2.28
## Proactive Claw v1.2.28 โ Changelog
- Added OpenClaw metadata to SKILL.md for improved compatibility and transparency.
- No functional or behavioral changes to core skill logic.
- Previous requirements and install fields are now structured under a metadata section.
- Documentation and description remain unchanged.
v1.2.27
Summary: Removes all optional third-party integrations for a privacy-first, standalone release.
- All third-party integrations (Telegram, Notion, LLM/cloud, Clawhub) removed; skill is now Google Calendar or Nextcloud CalDAV only.
- Cleaned up configurationโno feature flags or options for external integrations remain.
- Simplified documentation and install instructions to reflect only calendar access and core features.
- Removed scripts and code paths related to optional integrations.
v1.2.25
proactive-claw v1.2.25
- Removed requirement for the GOOGLE_CREDENTIALS_JSON environment variable; credentials now provided solely via config files.
- Minor documentation updates and clarifications in SKILL.md.
- No functional or API changes.
v1.2.23
**Proactive Claw v1.2.23 โ Improved calendar write safety & install clarity**
- Added a runtime guard in `cal_backend.py`: all event write operations now use `_assert_write_allowed()` to prevent modifications to any calendar except the Actions calendar defined in `config.json`.
- Clarified install instructions and credential flow; the helper script for clawhub OAuth is now located at `scripts/optional/setup_clawhub_oauth.sh` and confirmed as fully optional/not recommended for most users.
- Updated `SKILL.md` and `config.example.json` to document new runtime protections and improve accuracy on install/catalog steps.
- Deprecated and removed the legacy `config.txt` and old OAuth helper script; use new optional directory for helper tools.
- The `setup.sh` and install flow now reference `uv pip install` (with fallback to pip3), improving install reliability for both Google and Nextcloud backends.
v1.2.20
proactive-claw v1.2.20
- Updated SKILL.md to reflect new version (1.2.20).
- No functional changes, code updates, or behavioral modifications in this release.
- Documentation remains focused on installation steps, credentials, and usage transparency.
v1.2.19
**Enhanced security transparency and OAuth handling, plus new clawhub.ai integration opt-in**
- Added SECURITY.md for clear disclosure of permissions, OAuth scope, and write protections
- setup.sh and install instructions now explicitly document Google OAuth scope and code-level safeguards: only the skillโs Actions calendar is writable, all other calendars strictly read-only by code policy
- Added scripts/setup_clawhub_oauth.sh for optional integration with clawhub.ai to fetch credentials.json; not invoked by default, requires explicit user action
- Install steps and side effects now clarify all package sources (PyPI only), privilege levels (no root/sudo ever), and directory write boundaries
- Minor improvements to SKILL.md instructions and wording for greater transparency and security clarity
v1.2.18
- Refreshed skill documentation for clarity and transparency: install steps, credential handling, and side effects are now fully explicit in the SKILL.md.
- Improved section on required/optional credentials, with clear tables and descriptions of both Google and Nextcloud backends.
- Detailed all install actions, including Python dependency sources (PyPI only) and background daemon setup, with full steps shown.
- Enhanced security and privacy disclosures, reinforcing that no credentials are uploaded and all actions are local and inspectable.
- No changes to functionality or APIsโdocumentation only.
v1.2.17
- Clarified credential requirements in a new summary table at the top (frontmatter and documentation).
- Updated environment variable from GOOGLE_OAUTH_CREDENTIALS_JSON to GOOGLE_CREDENTIALS_JSON.
- Streamlined and shortened the description and requirements for easier onboarding.
- Improved separation of required vs optional credentials for Google and Nextcloud backends.
- No functional or code changes; documentation and metadata only.
v1.2.16
- Added primaryEnv: GOOGLE_OAUTH_CREDENTIALS_JSON and documented required environment variables/config files.
- Expanded "requires" to include env and config details.
- Clarified setup requirements for Google OAuth and Nextcloud credentials in the SKILL.md.
- No changes to logic, features, or side effectsโdocumentation/adoption improvements only.
Metadata
Frequently Asked Questions
What is Proactive Claw?
๐ฆ The most powerful proactive engine for OpenClaw. Your personal assistant that learns from you and helps you be more productive. It is an AI Agent Skill for Claude Code / OpenClaw, with 1387 downloads so far.
How do I install Proactive Claw?
Run "/install proactive-claw" in the OpenClaw or Claude Code chat to install it in one step โ no extra setup required.
Is Proactive Claw free?
Yes, Proactive Claw is completely free (open-source). You can download, install and use it at no cost.
Which platforms does Proactive Claw support?
Proactive Claw is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created Proactive Claw?
It is built and maintained by Dawid Piaskowski (@googlarz); the current version is v1.2.41.
More Skills