The Agent Skills Directory

[COMMAND_EXECUTION]: The skill requires the execution of shell commands for testing, linting, and formatting (e.g., pnpm test, node --test). These commands are essential for the TDD workflow and are typically configured on a per-project basis within the agent's environment.\n- [PROMPT_INJECTION]: The skill recommends a 'Test-Driven Prompting' (TDP) pattern in SKILL.md where raw stdout/stderr from a test runner is captured and injected directly into a sub-agent's prompt. This creates a surface for indirect prompt injection if an attacker can control the test output (e.g., through failing test descriptions or specific error messages).\n
Ingestion points: The testOutput variable used to construct spawn prompts for the developer sub-agent.\n
Boundary markers: Markdown code blocks (triple backticks) are used to wrap the injected output.\n
Capability inventory: The skill utilizes tools such as Bash, Write, Edit, and Read across its instructions and hooks.\n
Sanitization: No sanitization or 'ignore-instructions' delimiters are applied to the raw test output before it is interpolated into the prompt.\n- [EXTERNAL_DOWNLOADS]: The skill refers to several external packages and tools including pnpm, stryker, msw, and fast-check. These are downloaded from official and well-known registries (e.g., npm) during standard development operations. No suspicious or unknown remote sources were detected.

tdd