accessibility-regression-detector
Pass
Audited by Gen Agent Trust Hub on Sep 26, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONNO_CODE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill defines detection modes that rely on processing external, untrusted data such as accessibility audit reports and Git history, creating a surface for potential indirect prompt injection attacks.
- Ingestion points: Git history analysis and audit report comparison (SKILL.md).
- Boundary markers: Absent. The instructions do not define delimiters or specific 'ignore instructions' warnings for the agent to use when evaluating external data.
- Capability inventory: None. The skill does not include any scripts, subprocess calls, or specific tool configurations.
- Sanitization: Absent. The instructions provide no guidance for sanitizing or validating the input data before processing.
- [NO_CODE]: The skill consists entirely of instructional markdown and does not package any executable scripts, binaries, or source code.
Audit Metadata