interface-audit
Pass
Audited by Gen Agent Trust Hub on Sep 11, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONPERSISTENCE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests UI source code files which are untrusted external data that could potentially contain malicious instructions designed to subvert the agent's audit logic.
- Ingestion points: The skill reads files including components, pages, screens, templates, and HTML files (Phase 2).
- Boundary markers: The skill does not implement explicit delimiters or "ignore instructions" wrappers when processing the content of these files.
- Capability inventory: The skill has the ability to execute shell commands (lint/test/build) and perform file-write operations (auto-fixes) based on its analysis of the ingested data.
- Sanitization: No evidence of content sanitization or instruction filtering was found.
- [COMMAND_EXECUTION]: The skill executes the project's existing lint, test, and build commands (Phase 5) to verify the integrity of auto-fixes. This execution is limited to pre-existing tools defined within the analyzed repository's configuration.
- [PERSISTENCE]: The skill maintains a local learning log by appending metadata and findings to a hidden file in the user's home directory (
~/.claude/skills/interface-audit/LEARNINGS.md) to track history across sessions.
Audit Metadata