validating-accessibility
Pass
Audited by Gen Agent Trust Hub on Feb 28, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill describes a 'MANDATORY' workflow using TeamCreate to spawn specialized agents within the fuse-ai-pilot namespace. These agents are identified as vendor resources from the skill's author (fusengine), representing normal multi-agent orchestration rather than a malicious bypass attempt.
- [SAFE]: The skill implements tools for auditing and editing codebase files. While reading external files creates a surface for indirect prompt injection, this capability is essential for the skill's primary goal of accessibility validation and no malicious behavior was detected.
- [SAFE]: All external references are local markdown files containing verified UX and accessibility standards. No obfuscation, data exfiltration, or dangerous command execution patterns are present in the provided files.
Audit Metadata