opus_4_5_migration
Pass
Audited by Gen Agent Trust Hub on Feb 27, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill references documentation and best practices from trusted and well-known sources, including Anthropic's official API documentation and the Prompting Guide website.
- [COMMAND_EXECUTION]: The migration workflow involves searching for and updating model strings within a codebase. These file system operations are standard for a migration utility and are limited to the skill's documented scope.
- [PROMPT_INJECTION]: The skill operates on untrusted data (user codebases and prompts) to perform modifications. It appropriately addresses the risk of indirect prompt injection by recommending the use of XML tags (e.g., <code_guidelines>, <tool_usage>) to provide clear boundaries for integrated snippets.
- Ingestion points: Codebase files containing model strings and API call configurations.
- Boundary markers: Explicit instructions to use XML tags and match the existing structure of processed prompts to maintain instruction isolation.
- Capability inventory: File read/write operations for updating model strings and prompt contents.
- Sanitization: Relies on structural separation (XML) to prevent embedded instructions in migrated code from being misinterpreted as agent commands.
Audit Metadata