dashboard
Warn
Audited by Gen Agent Trust Hub on Feb 23, 2026
Risk Level: MEDIUMDATA_EXFILTRATIONCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [DATA_EXFILTRATION]: The skill documentation references absolute file paths on the author's local system.
- Evidence: References to
/home/mwguerra/projects/mwguerra/claude-code-plugins/filament-specialist/skills/docs/references/disclose specific directory structures and user account names. - [COMMAND_EXECUTION]: The skill provides instructions and templates for generating and writing executable PHP and Blade files to the filesystem.
- Evidence: Templates for
App\Filament\__PANEL__\Pages\__PAGE_CLASS__and corresponding Blade views intended to be written toresources/views/filament/. - [PROMPT_INJECTION]: The skill creates a surface for Indirect Prompt Injection or XSS by generating templates that disable HTML sanitization.
- Ingestion points: User-provided
messagestrings in the tab schema contract are directly interpolated into templates. - Boundary markers: No boundary markers or instructions to sanitize user input are present for the generated code.
- Capability inventory: The skill has the capability to write PHP and Blade files to the application structure.
- Sanitization: Explicitly missing; the template uses
{!! $activeTabData['message'] !!}which renders unescaped HTML, creating a vulnerability if the dashboard content comes from untrusted sources.
Audit Metadata