The Agent Skills Directory

[SAFE]: The skill provides instructions for the agent to act as a static code analyzer for documentation. No malicious patterns such as prompt injection, unauthorized data access, or remote code execution were found.- [SAFE]: The tool 'microsoft_code_sample_search' referenced in the instructions is a specialized search utility for verifying API status, which is consistent with the skill's purpose and the author (MicrosoftDocs). No excessive or dangerous capabilities are requested.- [PROMPT_INJECTION]: The skill processes user-provided code samples, which serves as an ingestion point for untrusted data (Indirect Prompt Injection surface). However, the instructions specify that the agent should 'Ignore prose text outside of code blocks' and 'Report ONLY definite, verifiable errors', which significantly reduces the risk of following embedded instructions. Capability inventory is limited to a specific search tool, and no file-system or shell-execution capabilities are exposed.

code-expert