The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to process external, untrusted artifacts such as task specifications, source code diffs, and test results.
Ingestion points: The agent is instructed to read and analyze third-party specifications, code, and test evidence.
Boundary markers: The skill mandates the use of a requirement matrix and structured reporting to keep analyzed data separate from the agent's logic.
Capability inventory: The workflow involves standard development tools like git, npm, and xcodebuild to verify implementation state.
Sanitization: Instructions specifically emphasize redacting operational details and avoiding the movement of secrets to unsafe environments.
[COMMAND_EXECUTION]: The skill references standard development and diagnostic utilities as part of a verification workflow.
Evidence: Mentions using git status, npm test, and xcodebuild for capturing compliance evidence.
Evidence: Provides legitimate troubleshooting instructions for macOS/iOS development environments, such as using xcode-select to configure the active toolchain.

specification-compliance-review