skills/thomasrohde/agent0/write-a0/Gen Agent Trust Hub

write-a0

Pass

Audited by Gen Agent Trust Hub on Jun 22, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill serves as a technical reference for the A0 language, which implements a 'deny-by-default' security model via capability declarations (e.g., cap { fs.read: true }). This requires scripts to explicitly request permission for sensitive operations.\n- [SAFE]: The documentation introduces resource 'budgets' (e.g., timeMs, maxToolCalls), which allow for limiting the execution impact of scripts and preventing runaway processes.\n- [SAFE]: All included example programs and idiomatic patterns demonstrate benign use cases, such as JSON data transformation, workspace discovery, and standard API interaction using well-known placeholder services (e.g., jsonplaceholder.typicode.com).\n- [SAFE]: No evidence of prompt injection, data exfiltration, obfuscation, or unauthorized command execution was found in any of the skill's instructional or reference files.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 22, 2026, 02:58 AM
Security Audit — agent-trust-hub — write-a0