cfe-patch-method
Pass
Audited by Gen Agent Trust Hub on Oct 27, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill performs legitimate development automation tasks for 1C Enterprise extensions. It reads and writes source code files (.bsl and .xml) based on explicit user-provided paths.
- [COMMAND_EXECUTION]: The skill executes a local PowerShell or Python script included in the skill package. These operations are confined to the local filesystem and act on paths provided as arguments.
- [INDIRECT_PROMPT_INJECTION]: The skill reads external project files. It uses deterministic parsing and regex-based extraction to identify code blocks, minimizing the risk of instructions in the source code influencing the agent's behavior. Ingestion points: modules and XML files; Boundary markers: regex parsing; Capabilities: file write; Sanitization: string escaping.
Audit Metadata