The Agent Skills Directory

Standard Code Extraction and Repository Workflow Alignment: The skill automates standard tasks associated with PR review and testing in the target microsoft/aspire repository. It provides operational workflows for checking out the code, fetching the correct testing tools, running test applications, and generating reports.
External Scripts from Vendor Organization (Safe): The skill utilizes curl and Invoke-RestMethod (irm) to pull specific helper scripts (get-aspire-cli-pr.sh and get-aspire-cli-pr.ps1) directly from the official microsoft/aspire GitHub repository. Under the [TRUST-SCOPE-RULE] and author context rules, these downloads are classified as safe because they belong to the trusted vendor's own verified infrastructure.
Secure Credential Handling Considerations: The script provides explicit guidance not to override environment variables like HOME or APPDATA, which preserves the system's standard, native authentication configuration via the official GitHub CLI (gh). This minimizes risks related to token exposure and complies with recommended secret-handling practices.
Isolated Test Execution Environments: The instructions mandate the use of temporary directories (mktemp / $env:TEMP) or repository-local ephemeral containers for executing tests. This isolation approach follows best practices to ensure that local or build environments are not modified unexpectedly.
Indirect Prompt Injection Surface Evaluation: The skill ingests pull request numbers, body text, or patch lists using the GitHub CLI. It has specific safety guidelines embedded (such as prompt-suppressing options like --suppress-agent-init and explicit non-interactive flags) which mitigate potential input confusion surfaces when processing PR contents.

pr-testing