The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes a wide array of system-level test runners and build tools, including Maven, Gradle, Pytest, Dotnet, Jest, and Go. It also executes specific Python utility scripts from a plugin root directory variable.
[INDIRECT_PROMPT_INJECTION]: The skill ingests and processes content from project-specific files such as DELTA_CATALOG.md, BUSINESS_RULES.md, and INTENT.md to guide its test generation and verification logic.
Ingestion points: Content is read from DELTA_CATALOG.md, BUSINESS_RULES.md, INTENT.md, cases.json, and JUnit-style XML test reports.
Boundary markers: Absent. There are no explicit delimiters or instructions provided to the agent to treat the content of these files as untrusted data or to ignore embedded instructions.
Capability inventory: The agent has the capability to execute shell commands, run test suites, and write to the local file system (specifically the analysis/ and fresh/ directories).
Sanitization: Absent. No explicit sanitization or validation of the input file content is performed before it is used to influence the agent's logic for creating new test inputs or generating verification results.

modernize-verify