The Agent Skills Directory

[Remote Code Execution]: The skill downloads and executes setup scripts and binaries from remote URLs associated with the Microsoft organization.
Evidence: curl -fsSL https://raw.githubusercontent.com/microsoft/aspire/main/eng/scripts/get-aspire-cli-pr.sh | bash and irm ... | iex in SKILL.md.
Context: These operations are part of the standard installation process for testing PR-specific builds of the Microsoft Aspire CLI and originate from official repositories.
[Indirect Prompt Injection Surface]: The skill analyzes Pull Request diffs and comments to propose test scenarios, which could potentially contain instructions intended to influence the agent's behavior.
Ingestion points: PR data is ingested via gh pr view and gh pr diff as described in SKILL.md.
Boundary markers: The skill uses jq to isolate specific comments but lacks explicit separators or "ignore instructions" when processing the full diff or comment text.
Capability inventory: The skill can execute shell commands, run the downloaded CLI, and manage local files.
Sanitization: No automated sanitization of PR content is performed, but the skill mitigates this by requiring explicit user confirmation via the ask_user tool before executing any proposed scenarios.
[Command Execution]: The skill makes extensive use of shell commands including gh, bash, powershell, and docker to prepare environments and execute tests.
Context: This is routine for a development-focused skill intended to automate software testing and validation tasks.

pr-testing