ci-accessibility
Pass
Audited by Gen Agent Trust Hub on Sep 26, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill demonstrates a vulnerability surface by processing untrusted data. 1. Ingestion points: HTML file discovery and scanning from pull request changed files and existing CI configuration files as described in SKILL.md (Phase 1 and Phase 3). 2. Boundary markers: Absent; the skill lacks explicit instructions to treat embedded data as untrusted or to use markers to separate it from instructions. 3. Capability inventory: The skill generates configurations that involve shell command execution (axe-core CLI) and performs file writes for baseline management in SKILL.md (Phase 3 and Phase 4). 4. Sanitization: No sanitization or filtering logic is specified for the external HTML content before it is processed by scanning tools.
Audit Metadata