The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides instructions for running standard software testing commands to verify implementation changes.
Evidence: SKILL.md and references/EXAMPLES.md contain examples of executing test runners such as npm test, pytest, go test ./..., and cargo test using the Bash tool.
[INDIRECT_PROMPT_INJECTION]: The skill's workflow involves reading and executing tests against project source code, which creates a theoretical surface for data-driven injection.
Ingestion points: Project source code and test files (e.g., cart.test.ts, test_dashboard.py) read by the agent to perform the TDD loop.
Boundary markers: The skill does not define specific isolation markers or instructions to disregard embedded commands in the code being tested.
Capability inventory: The agent has access to the Bash tool to run package managers and test suites across multiple languages.
Sanitization: No sanitization is performed on the code content before it is processed or used in test execution commands.
[DYNAMIC_EXECUTION]: The documentation includes educational content on advanced testing strategies that involve runtime environment modifications.
Evidence: references/LEGACY_CODE.md describes "Link seams" and mentions LD_PRELOAD as a mechanism for swapping implementations at build or import time. This is presented as a strategy for testing legacy systems rather than an instruction for malicious behavior.

tdd-workflow