The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data in the form of 'Requirement descriptions' and 'Relevant code file paths' to generate tests. While this presents an attack surface where malicious input could influence the generated code, the skill mitigates this by restricting the sub-agent's permissions (read-only on source files, writable only on test files).
Ingestion points: SKILL.md Step 1 identifies 'Requirement description' and 'Relevant code file paths' as primary inputs.
Boundary markers: Not explicitly defined for the external requirement strings.
Capability inventory: The skill uses Read, Grep, Glob, and Write (limited to test files), and requires the execution of generated tests (Step 4).
Sanitization: No explicit sanitization of input requirements is mentioned, though the skill relies on the underlying platform's agent isolation.
[COMMAND_EXECUTION]: The skill directs the user or agent to 'Run the tests' to validate they fail as expected (Step 4). This involves executing generated code. While execution of generated code is high-risk in some contexts, it is the fundamental purpose of a TDD/Testing skill and is confined to the project's test environment.

test-designer