AI Newsletter Digest improvements: fixed QP soft line break decoding, URL extraction, and content cleaning

This commit is contained in:
Krilly
2026-03-04 13:29:22 +00:00
parent 29a98137a7
commit 57dd294675
13706 changed files with 2114953 additions and 237629 deletions

View File

@@ -0,0 +1,57 @@
# OpenClaw Advanced Tools Reference
Reference normalized against:
- `https://docs.openclaw.ai/tools`
- `https://docs.openclaw.ai/tools/browser`
- `https://docs.openclaw.ai/automation/cron-jobs`
- `https://docs.openclaw.ai/plugins`
- `https://docs.openclaw.ai/cli/gateway`
Last verified: 2026-02-17.
## Safety First
This file documents advanced capabilities for operator awareness.
It does not grant blanket authorization to execute privileged or high-risk operations.
See `references/security-policy.md`.
## Gateway API Utility
Use gateway API calls directly from CLI:
- `openclaw gateway call <path>`
Typical patterns:
- health checks
- custom automation hooks
- debugging specific gateway endpoints
## Managed Browser (High-risk)
- `openclaw browser start`
- `openclaw browser open <url>`
- `openclaw browser screenshot [--full-page]`
- `openclaw browser snapshot --format aria`
- `openclaw browser stop`
Requires explicit approval and wrapper opt-in.
## Cron Automation (High-risk)
- `openclaw cron list`
- `openclaw cron add ...`
- `openclaw cron run <jobId>`
- `openclaw cron remove <jobId>`
Creation, run, and delete operations require explicit approval.
## Plugins (High-risk)
- `openclaw plugins list`
- `openclaw plugins install <path-or-url>`
- `openclaw plugins enable <name>`
- `openclaw plugins disable <name>`
Install and enable operations should be limited to trusted sources.
## Upstream Runtime Capabilities (Awareness)
Depending on OpenClaw runtime configuration, upstream may expose:
- `exec` style arbitrary command execution
- elevated permission workflows
- sub-agent delegation
Treat these as privileged features with explicit, per-action approval.

View File

@@ -0,0 +1,62 @@
# OpenClaw Full CLI Reference
Reference normalized against `https://docs.openclaw.ai/cli`.
Last verified: 2026-02-17.
Safety note:
- Commands listed here reflect upstream CLI surface.
- Local wrapper policy may block high-risk groups unless `OPENCLAW_WRAPPER_ALLOW_RISKY=1`.
- See `references/security-policy.md`.
## Core Commands
- `openclaw onboard`: Run onboarding wizard.
- `openclaw doctor`: Validate local install and health.
- `openclaw status`: Show global OpenClaw status.
- `openclaw version`: Print CLI version.
- `openclaw dashboard [--no-open]`: Open or print dashboard URL.
- `openclaw config [--list|--get <key>|--set <key=value>]`: Inspect or modify config values.
- `openclaw update`: Update CLI to latest stable build.
- `openclaw uninstall`: Remove CLI and optionally data.
## Gateway Commands
- `openclaw gateway`: Start gateway in foreground.
- `openclaw gateway status`: Show gateway runtime status.
- `openclaw gateway health`: Check gateway health endpoint.
- `openclaw gateway restart`: Restart managed gateway process.
- `openclaw gateway stop`: Stop managed gateway process.
- `openclaw gateway install`: Install gateway background service.
- `openclaw gateway uninstall`: Remove gateway background service.
- `openclaw gateway call <path> [--method ...] [--data ...]`: Call gateway APIs.
## Channels and Pairing
- `openclaw channels list`: List configured channels.
- `openclaw channels login --channel <name>`: Authenticate a channel.
- `openclaw channels logout --channel <name>`: Disconnect a channel.
- `openclaw pairing`: Show pairing flow for mobile nodes (high-risk).
## Models
- `openclaw models list`: Show available models.
- `openclaw models set <model>`: Set default model.
- `openclaw models auth add --provider <openai|anthropic>`: Add model provider auth.
- `openclaw models auth list`: Show provider auth entries.
- `openclaw models auth remove --provider <name>`: Remove provider auth.
- `openclaw models aliases list`: List alias map.
- `openclaw models aliases add <alias> <model>`: Add alias.
- `openclaw models aliases remove <alias>`: Remove alias.
- `openclaw models scan`: Discover local models.
## Automation
- `openclaw cron list`: List cron jobs.
- `openclaw cron add ...`: Create cron job (high-risk).
- `openclaw cron run <jobId>`: Run job immediately (high-risk).
- `openclaw cron remove <jobId>`: Delete job (high-risk).
## Browser and Plugins
- `openclaw browser status|start|stop`: Manage browser runtime (high-risk).
- `openclaw browser open <url>`: Navigate browser (high-risk).
- `openclaw browser screenshot [--full-page]`: Capture screenshot.
- `openclaw browser snapshot --format aria`: Capture structured page snapshot.
- `openclaw plugins list`: List plugins.
- `openclaw plugins install <path-or-url>`: Install plugin (high-risk).
- `openclaw plugins enable <name>`: Enable plugin (high-risk).
- `openclaw plugins disable <name>`: Disable plugin.

View File

@@ -0,0 +1,64 @@
# OpenClaw Configuration Reference
Reference normalized against:
- `https://docs.openclaw.ai/cli/config`
- `https://docs.openclaw.ai/gateway/configuration`
Last verified: 2026-02-17.
## Config File Location
Default state directory:
- `~/.openclaw`
Default config file:
- `~/.openclaw/openclaw.json5`
## CLI-first Config Management
Use CLI first to avoid manual syntax errors:
- `openclaw config --list`
- `openclaw config --get gateway.bind`
- `openclaw config --set gateway.bind=127.0.0.1`
- `openclaw config --set gateway.port=18789`
## Minimal Config Example
```json5
{
gateway: {
bind: "127.0.0.1",
port: 18789,
auth: {
token: "replace-with-strong-token"
}
},
channels: {
whatsapp: {
allowFrom: ["+1234567890"],
groups: {
"*": { requireMention: true }
}
}
},
agents: {
defaults: {
workspace: "~/.openclaw/workspace"
}
},
models: {
default: "claude-3-5-sonnet-latest"
}
}
```
## High-impact Keys
- `gateway.bind`: Interface binding. Keep `127.0.0.1` unless remote access is required.
- `gateway.port`: Gateway port (default `18789`).
- `gateway.auth.token`: Required when binding beyond loopback.
- `channels.*`: Channel-specific policy and auth settings.
- `agents.defaults.workspace`: Base workspace for agent tasks.
- `models.default`: Default model used by agents.
## Environment Variables
- `OPENCLAW_CONFIG_PATH`: Override config file path.
- `OPENCLAW_STATE_DIR`: Override state directory.
- `OPENCLAW_HOME`: Override OpenClaw home directory.

View File

@@ -0,0 +1,43 @@
# OpenClaw Deployment and Operations
Reference normalized against:
- `https://docs.openclaw.ai/install`
- `https://docs.openclaw.ai/install/docker`
- `https://docs.openclaw.ai/install/updating`
- `https://docs.openclaw.ai/install/uninstall`
Last verified: 2026-02-17.
## Local Install and Onboarding
- Install CLI via official install flow from docs.
- Run `openclaw onboard` to initialize and connect providers/channels.
- Validate with `openclaw doctor`.
## Docker Deployment
- Use the official Docker guide under `/install/docker`.
- Typical lifecycle:
- `docker compose up -d`
- `docker compose logs -f`
- `docker compose down`
- Keep OpenClaw state mounted on persistent volume.
## Service Lifecycle
- `openclaw gateway install`: Install background service.
- `openclaw gateway restart`: Restart service.
- `openclaw gateway status`: Check runtime status.
- `openclaw gateway uninstall`: Remove service.
## Updating and Rollback
- Update CLI: `openclaw update`
- If rollback is needed, reinstall pinned version using the official package path described in docs.
- Re-run `openclaw doctor` after updates.
## Uninstall
- Use `openclaw uninstall` for full removal flow.
- Confirm whether state data should be preserved or removed.
## Production Safety Checklist
- Set strong `gateway.auth.token`.
- Avoid public bind unless required.
- Use VPN/Tailscale or private network for remote access.
- Monitor with `openclaw gateway health`.

View File

@@ -0,0 +1,41 @@
# OpenClaw Documentation Hubs
Primary docs root:
- https://docs.openclaw.ai/
Last verified: 2026-02-17.
## Start and Install
- https://docs.openclaw.ai/start/getting-started
- https://docs.openclaw.ai/start/onboarding
- https://docs.openclaw.ai/install
- https://docs.openclaw.ai/install/docker
- https://docs.openclaw.ai/install/updating
- https://docs.openclaw.ai/install/migrating
- https://docs.openclaw.ai/install/uninstall
## CLI and Gateway
- https://docs.openclaw.ai/cli
- https://docs.openclaw.ai/cli/config
- https://docs.openclaw.ai/cli/gateway
- https://docs.openclaw.ai/cli/channels
- https://docs.openclaw.ai/cli/models
- https://docs.openclaw.ai/cli/cron
- https://docs.openclaw.ai/cli/browser
- https://docs.openclaw.ai/cli/plugins
- https://docs.openclaw.ai/cli/pairing
## Platform and Architecture
- https://docs.openclaw.ai/platforms
- https://docs.openclaw.ai/platforms/windows
- https://docs.openclaw.ai/platforms/macos
- https://docs.openclaw.ai/platforms/linux
- https://docs.openclaw.ai/gateway/configuration
- https://docs.openclaw.ai/gateway/network-model
- https://docs.openclaw.ai/gateway/multiple-gateways
## Tools and Automation
- https://docs.openclaw.ai/tools
- https://docs.openclaw.ai/tools/browser
- https://docs.openclaw.ai/automation/cron-jobs
- https://docs.openclaw.ai/plugins

View File

@@ -0,0 +1,34 @@
# OpenClaw Nodes and Platforms
Reference normalized against:
- `https://docs.openclaw.ai/platforms`
- `https://docs.openclaw.ai/platforms/windows`
- `https://docs.openclaw.ai/platforms/macos`
- `https://docs.openclaw.ai/platforms/linux`
- `https://docs.openclaw.ai/cli/pairing`
Last verified: 2026-02-17.
## Windows (WSL2)
- Use a Linux distro on WSL2 for stable daemon behavior.
- Keep gateway on loopback unless remote access is intentionally configured.
- For remote node access, prefer private networking (for example Tailscale).
## macOS
- Use macOS-specific platform guidance from docs for companion integrations.
- Ensure required local permissions are granted before enabling device features.
## Linux
- Install gateway service for persistent runtime.
- Validate prerequisites for browser tooling when using managed browser commands.
## Mobile Nodes and Pairing
- Run `openclaw pairing` to start pairing flow.
- Complete pairing from OpenClaw mobile app.
- Verify node availability from dashboard or status commands.
## Node Security Baseline
- Require gateway token for non-loopback bind.
- Restrict who can message/control agent via channel policy.
- Review permissions before enabling camera, microphone, or location features.
- Treat pairing and sensor access as high-risk actions requiring explicit approval.

View File

@@ -0,0 +1,47 @@
# OpenClaw Prerequisites and Capability Boundaries
Last verified: 2026-02-17.
Source of truth: `https://docs.openclaw.ai`.
## Required
- `openclaw` CLI installed and accessible in `PATH`
## Optional or Feature-dependent
- Node.js and npm: install/update workflows
- Playwright OS dependencies: managed browser features
- `imsg` on macOS: iMessage channel integrations
- Tailscale or equivalent private networking: remote/mobile node access
- Docker Engine + Compose: containerized deployments
- Nix: flake-based environments
## Runtime Environment Variables
- `OPENCLAW_CONFIG_PATH`
- `OPENCLAW_STATE_DIR`
- `OPENCLAW_HOME`
## Capability Boundaries
Low-risk by default:
- status checks
- diagnostics
- read-focused configuration inspection
High-risk (explicit approval required):
- arbitrary shell execution features in upstream runtime
- elevated mode
- sub-agent execution with inherited environment
- browser automation, plugin install/enable, cron mutation
- device pairing and sensor-enabled features
Wrapper-level control:
- `OPENCLAW_WRAPPER_ALLOW_RISKY=1` is required for risky command groups in `scripts/openclaw.sh`.
## What This Repo Provides
- Local docs and references
- A shell wrapper (`scripts/openclaw.sh`) that routes commands to `openclaw`
## What This Repo Does Not Provide
- Binary installation of OpenClaw itself
- Automatic OS package provisioning
- Host-level network/VPN setup
- Mobile platform permissions and device configuration
- Automatic approval for privileged actions

View File

@@ -0,0 +1,34 @@
# OpenClaw Security Policy
Last verified: 2026-02-17.
## Purpose Constraint
This skill exists to operate OpenClaw safely as a CLI wrapper plus documentation aid.
It is not a blanket authorization for privileged or autonomous actions.
## Approval Model
- Low-risk read operations may run by default.
- High-risk operations require explicit user approval for each action.
- Do not chain high-risk actions into unattended workflows by default.
## High-risk Categories
- Shell execution features (`exec`) that can run arbitrary commands
- Elevated privilege flows
- Sub-agent delegation with inherited environment/context
- Plugin install from external sources
- Cron add/remove/force-run operations
- Browser automation on arbitrary remote sites
- Device pairing and sensor access (camera/audio/location)
## Required Controls
- Principle of least privilege
- Explicit, contextual consent before each high-risk step
- Prefer read-only checks before any mutating action
- Use trusted plugin sources only
- Keep gateway bound to loopback unless remote access is intentional
## Wrapper Enforcement
`bash scripts/openclaw.sh` blocks high-risk command groups unless:
- `OPENCLAW_WRAPPER_ALLOW_RISKY=1`
This opt-in is session-scoped and should be set only when required.