sillytavern-embedded-ui
Pass
Audited by Gen Agent Trust Hub on Oct 23, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from various sources but implements defenses: (1) Ingestion points: Data from messages, model output, and remote UI sources are identified in SKILL.md. (2) Boundary markers: The skill recommends using namespace IDs and clear content boundaries in SKILL.md. (3) Capability inventory: The skill interacts with project watchers in SKILL.md and uses Get-FileHash for integrity in references/sandbox-recipes.md. (4) Sanitization: SKILL.md mandates the use of trusted sanitizers, attribute escaping, and avoiding innerHTML.
- [COMMAND_EXECUTION]: The skill mentions using 'Get-FileHash' for integrity checks and 'sillytavern-card-pipeline' for running project watchers, which are standard development tasks but involve executing external tools.
Audit Metadata