Is swiftui-ui-patterns safe?

https://clawhub.ai/Dimillian/swiftui-ui-patterns

96
SAFE

This is a benign documentation-only skill containing SwiftUI UI patterns, component architecture guidance, and Swift code examples. It has no executable code, no network access, no package.json scripts, no git hooks, and no data exfiltration vectors. The skill's only agent-directing instructions are standard developer workflow suggestions (search codebase with ripgrep, read source files).

Category Scores

Prompt Injection 95/100 · 30%
Data Exfiltration 100/100 · 25%
Code Execution 100/100 · 20%
Clone Behavior 95/100 · 10%
Canary Integrity 100/100 · 10%
Behavioral Reasoning 85/100 · 5%

Findings (2)

INFO Benign agent workflow instructions -5 ▶

The skill instructs the agent to use 'rg' (ripgrep) to search the codebase and read SwiftUI view files. These are standard developer workflow suggestions that cause the agent to execute shell commands and read files, but only within the user's own project context.

INFO Audit harness reads sensitive paths during startup -5 ▶

The filesystem monitor recorded access to .aws/credentials, .env, and .openclaw config files. Investigation confirms these are from the audit harness (openclaw agent) startup sequence, not from the skill under test. The skill was installed after these accesses occurred.