← Back to Skills Marketplace
172
Downloads
0
Stars
0
Active Installs
3
Versions
Install in OpenClaw
/install wander
Description
Monitor any async task that takes time and needs completion notification — CI, builds, deploys, releases. Use when the user triggers any long-running async t...
Usage Guidance
This package is plausible for watching GitHub Actions, but the documentation and actual runtime steps don't match the published metadata. Before installing: (1) review the upstream repository (https://github.com/ERerGB/wander) and inspect the watch-workflow*.sh scripts for any network calls or unexpected behavior; (2) be aware the agent will use your locally-authenticated gh CLI (so it can access private repos and Action logs) — check the GH auth scopes; (3) don't add PATH/aliases or run chmod +x until you trust the code; (4) if you don't want autonomous background watchers, instruct the agent to never auto-start Wander or to require explicit approval each time; (5) consider running the scripts in an isolated environment (container or throwaway VM) first. If you want, I can fetch and summarize the watch scripts so you can review the exact commands they run (I will not run them).
Capability Analysis
Type: OpenClaw Skill
Name: wander
Version: 1.0.6
The 'wander' skill is a legitimate utility designed to monitor GitHub Actions workflows and notify users upon completion. The SKILL.md file provides clear instructions for the AI agent to automatically trigger background monitoring scripts (e.g., watch-workflow-bg.sh) and poll for results using the GitHub CLI (gh). While the instructions include 'mandatory' auto-triggering of background processes, this behavior is transparently communicated to the user and is strictly aligned with the stated purpose of the tool. No evidence of malicious intent, data exfiltration, or unauthorized access was found in the provided code or instructions.
Capability Assessment
Purpose & Capability
The skill claims to be an async-task monitor for GitHub Actions (coherent) but the package metadata lists no requirements while the SKILL.md clearly requires the gh CLI, jq, and a local checkout of the Wander repo (WANDER_HOME). The metadata omission (no declared binaries or primary credential) is inconsistent with the runtime needs.
Instruction Scope
SKILL.md instructs the agent to read repository files (.workflows.yml and .github/workflows/), run downloaded shell scripts (watch-workflow*.sh), poll GH Actions via gh, and automatically start background monitoring after various triggers. Reading repo workflow files and calling gh is expected for this purpose, but the mandatory auto-triggering and background process start are broad and could be surprising if not explicitly consented to.
Install Mechanism
There is no formal install spec in the registry, but SKILL.md directs the user/agent to git clone https://github.com/ERerGB/wander.git and chmod +x the scripts. That is a network download and will place and execute code on disk; the source is GitHub (not a random pastebin), but the repository and its scripts should be inspected before running.
Credentials
The published requirements list no env vars or credentials, yet the instructions require an authenticated gh CLI (implicitly using the user's GitHub credentials), WANDER_HOME config and PATH/alias modifications, and writing logs to ~/.wander_logs/. Requesting GH access is proportionate to monitoring Actions, but the missing declaration of this dependency and the requirement to modify shell rc are mismatches.
Persistence & Privilege
The SKILL.md asks to persist scripts to a user directory, add PATH/aliases to shell rc, and run background/detached watchers that poll GH and write logs. These create persistent on-disk presence and autonomous background activity — acceptable for a watcher but potentially surprising and a privacy/availability consideration if enabled without explicit user consent.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install wander - After installation, invoke the skill by name or use
/wander - Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.6
CD: ff72575bc1c54494aabf8e4560f7c641d334df7e
v1.0.5
CD: c9eb8dd2ec2f022de8b3b6094fff0fdc06546118
v1.0.4
CD: 435e2524b71eb8195bd3b651930cb9ce70db4c38
Metadata
Frequently Asked Questions
What is Wander?
Monitor any async task that takes time and needs completion notification — CI, builds, deploys, releases. Use when the user triggers any long-running async t... It is an AI Agent Skill for Claude Code / OpenClaw, with 172 downloads so far.
How do I install Wander?
Run "/install wander" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is Wander free?
Yes, Wander is completely free, licensed under MIT-0. You can download, install and use it at no cost.
Which platforms does Wander support?
Wander is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created Wander?
It is built and maintained by ERerGB (@erergb); the current version is v1.0.6.
More Skills