run-apichief
Pass
Audited by Gen Agent Trust Hub on Sep 21, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- Command Execution: The skill invokes repository-specific scripts such as
restore.cmd,activate.ps1, andbuild.cmd, along with standard.NETbuild commands. These operations are expected for a tool designed to perform API baseline management and validation within a specific project environment. - Indirect Prompt Injection Surface: The skill identifies project folders and extracts metadata (like
AssemblyNameandTargetPath) from local.csprojfiles to construct its commands. While this is a standard automation pattern, it defines a surface where the contents of the local repository influence the execution parameters of the agent. - Ingestion points: The skill scans the
src/directory and executesmsbuildto retrieve project properties fromSKILL.mdinstructions. - Boundary markers: No explicit delimiters are used for the discovered file paths or properties.
- Capability inventory: The skill uses
dotnet build,msbuild, and script execution to process the repository. - Sanitization: The skill applies
.Trim()to command outputs before using them as variables in subsequent shell executions. - Dynamic Path Execution: The skill dynamically calculates paths for target assemblies and the
ApiChiefbinary at runtime using file system discovery and build system queries. This allows the skill to correctly target the appropriate build artifacts (e.g., specific TFMs) without hardcoded paths.
Audit Metadata