Setup & Installation
Or with the ClawHub CLI, for registry-managed skill folders outside a full OpenClaw workspace:
What This Skill Does
Compares two versions of an OpenClaw skill to surface security-relevant changes introduced between releases. Detects new network access, credential reads, code execution patterns, data exfiltration, and suspicious file additions. Produces a SAFE, REVIEW, or BLOCK recommendation based on what changed.
A static scanner checks a single version for known bad patterns, but the differ catches capabilities that are new between versions, which is where supply chain attacks hide.
When to use it
- Auditing a skill update before applying it from ClawHub
- Catching new network calls added in a minor version bump
- Detecting credential access that wasn't in the previous release
- Spotting new script files added alongside a claimed bug fix
- Blocking updates that silently introduce eval or exec patterns
Example Workflow
Here's how your AI assistant might use this skill in practice.
User asks: Is v1.1 of some-skill safe to update to?
- 1Download the new skill version to a temporary directory
- 2Run differ.py with --old pointing to the installed version and --new pointing to the download
- 3Parse output for newly detected capabilities across all changed files
- 4Map findings to a SAFE, REVIEW, or BLOCK recommendation
REVIEW — new outbound POST request detected in scripts/sync.py that did not exist in v1.0. Read the diff before updating.