The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses curl and curl.exe to interact with Azure DevOps REST APIs for project, repository, and security alert management.
[DATA_EXFILTRATION]: The skill accesses sensitive environment variables (AZURE_DEVOPS_PAT, AZURE_DEVOPS_TOKEN) for authentication. It interacts exclusively with well-known Azure DevOps domains (dev.azure.com and advsec.dev.azure.com) and includes explicit instructions to avoid printing, logging, or writing these credentials to files.
[INDIRECT_PROMPT_INJECTION]: The skill processes data from external sources that could be influenced by an attacker.
Ingestion points: Fetches security alert details, rule descriptions, and help messages from the Azure DevOps Advanced Security API (SKILL.md).
Boundary markers: The instructions mandate the use of tables for presenting results, which provides some structural separation, but no explicit instructions are provided to the agent to disregard potential instructions embedded within the fetched API data.
Capability inventory: The agent can execute curl commands and perform file operations in PowerShell environments (e.g., creating temporary JSON files for API payloads).
Sanitization: The skill does not specify any sanitization, filtering, or escaping for the tools[].rules[].description or helpMessage fields before they are presented to the user.

security-alert-review