claude-opus-4-5-migration
Pass
Audited by Gen Agent Trust Hub on Sep 30, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to search through and modify a user's codebase, which acts as an untrusted external data source. This creates a potential surface for indirect prompt injection if the codebase contains malicious instructions designed to influence the agent during the migration task.
- Ingestion points: The migration workflow requires the agent to "Search codebase for model strings and API calls" across all files (SKILL.md).
- Boundary markers: The skill suggests using XML tags (e.g.,
<frontend_aesthetics>,<code_guidelines>) when integrating new prompt snippets, providing some structural delimitation (references/prompt-snippets.md). - Capability inventory: To complete its task, the agent must employ file-reading, search, and file-writing tools to identify and update strings throughout the project.
- Sanitization: The instructions do not specify any validation or sanitization of the codebase content before the agent processes or interprets it.
Audit Metadata