The Agent Skills Directory

Prompt Injection (LOW): The skill exhibits a surface for indirect prompt injection. It analyzes external source code provided by the user, which may contain malicious instructions designed to manipulate the subagent's behavior during test generation.
Ingestion points: Source code paths provided in the /test:generate-tests command are ingested and analyzed as seen in references/generate-tests.md.
Boundary markers: No boundary markers (e.g., delimiters) or "ignore embedded instructions" warnings are defined in the delegation prompts to subagents.
Capability inventory: The subagents (test-automator and quality-engineer) possess Read, Grep, Glob, and Write capabilities to interact with the file system.
Sanitization: No sanitization or validation of the ingested code content is specified.
No Code (SAFE): This skill does not distribute any executable scripts or binary files, reducing the risk of direct local command execution or persistence.

test-generation