The Agent Skills Directory

Indirect Prompt Injection Surface: The skill facilitates the reading and analysis of external data sources, specifically project files and replayed build logs. These inputs represent a potential surface for indirect prompt injection, as malicious instructions could be embedded in code comments or build messages to influence the agent's behavior. * Ingestion points: Workspace project and solution discovery (SKILL.md) and diagnostic log files replayed from binary log artifacts (references/binlog-failure-analysis.md). * Boundary markers: The instructions do not define specific delimiters or instructions for the agent to treat replayed log content as untrusted data. * Capability inventory: The skill utilizes powerful CLI tools including dotnet build, dotnet msbuild, and dotnet publish, along with shell commands like Get-Content to process information. * Sanitization: There are no explicit sanitization or filtering steps for the content extracted from build logs before it is processed by the agent.
Sensitive Data Handling: The skill appropriately identifies that binary logs and replayed text logs can contain sensitive information such as credentials or environment variables. It includes explicit instructions for the user to review and redact these artifacts before sharing, which is a positive security practice (references/binlog-generation.md).

msbuild