The Agent Skills Directory

[COMMAND_EXECUTION]: The Python script scripts/investigate.py uses the subprocess module to execute external CLI tools, specifically the taskcluster and az (Azure) command-line interfaces.
[REMOTE_CODE_EXECUTION]: The skill provides a wrapper for executing arbitrary PowerShell scripts on remote Azure Virtual Machines via the az vm run-command invoke command. This is used to gather system information such as Windows build numbers, installed hotfixes, and worker configuration data.
[EXTERNAL_DOWNLOADS]: The documentation in SKILL.md describes workflows that use curl to fetch job data from treeherder.mozilla.org and recommends installing CLI prerequisites via Homebrew.
[PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by ingesting data from external APIs.
Ingestion points: scripts/investigate.py retrieves JSON responses from Taskcluster and Treeherder API endpoints.
Boundary markers: No explicit delimiters or instructions to ignore embedded commands are present in the prompt templates or data processing logic.
Capability inventory: The skill possesses the ability to execute local subprocesses and remote PowerShell commands on cloud instances.
Sanitization: Data is parsed using standard JSON libraries, but the content is not sanitized for potential malicious instructions before being presented to the agent context.

worker-image-investigation