The Agent Skills Directory

Indirect Prompt Injection Surface: The skill ingests codebase content to generate tests. If the processed code contains hidden instructions, the agent might follow them during the generation or execution phases.
Ingestion points: Reads source modules, test files, and project manifests to understand the codebase structure and conventions (SKILL.md, unit-test-generation.prompt.md).
Boundary markers: Absent; the instructions do not specify the use of clear delimiters or warnings to ignore embedded instructions within the ingested code files.
Capability inventory: The skill is capable of executing shell commands for building and testing, such as dotnet test, npm test, pytest, and git rev-parse (SKILL.md).
Sanitization: Absent; there is no explicit instruction to sanitize or validate the content of the files read from the environment.
Command Execution: The skill utilizes system commands to manage its internal state and validate generated tests.
Evidence: It uses git rev-parse to identify a non-staged directory for state management and invokes various language-specific test runners based on the project environment (SKILL.md).
Dynamic Execution: The skill generates executable test scripts and subsequently runs them to verify their correctness.
Evidence: It describes a multi-stage pipeline where it writes new test scripts (e.g., in C#, TypeScript, or Python) and executes them using the local test infrastructure (SKILL.md, unit-test-generation.prompt.md).

code-testing