skill-builder
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFE
Full Analysis
- [Indirect Prompt Injection] (LOW): The
SKILL.template.mddefines a 'Diagnostic Process' that ingests untrusted 'Writer' input, creating a potential ingestion surface. 1. Ingestion points:SKILL.template.md(Diagnostic Process). 2. Boundary markers: Absent. 3. Capability inventory: Local script execution (Deno) and filesystem persistence. 4. Sanitization: Absent. - [Command Execution] (SAFE): Examples in the template utilize the
deno run --allow-readcommand, which restricts script permissions to read-only access on the local filesystem, preventing unauthorized network or write operations. - [Data Exposure] (SAFE): Data access patterns are limited to local project directories (
../data/,context/). No hardcoded credentials, sensitive system paths, or non-whitelisted network operations were detected in the templates.
Audit Metadata