The Agent Skills Directory

Indirect Prompt Injection Surface: The skill is designed to ingest and summarize external test data (JSON, XML, and Markdown files) from a storage container. This creates a surface where instructions embedded in test logs could potentially influence agent behavior.
Ingestion points: Integration report blobs stored in the strdashboarddevveobvk Azure storage account (SKILL.md).
Boundary markers: The instructions do not specify the use of delimiters or specific 'ignore' instructions when processing the content of these files.
Capability inventory: The skill utilizes az storage blob download to fetch data and mcp_azure_mcp_storage_blob_get for discovery (SKILL.md).
Sanitization: There are no explicit requirements for sanitizing or escaping the content retrieved from the test logs before presentation to the user.
Command Execution: The skill utilizes the Azure CLI (az) to download files to the local system and requires executing shell commands (e.g., echo $env:TEMP) to resolve local environment variables. This is an intended capability for interacting with Azure resources but involves executing shell commands with parameters derived from storage metadata.
Local File Path Exposure: Instructions require the agent to generate vscode://file/ links pointing to the user's local temporary directory ($env:TEMP). This reveals internal filesystem paths in the final output, which is a consideration for data privacy in shared environments.

analyze-skill-issues