The Agent Skills Directory

[COMMAND_EXECUTION]: The skill is configured to use the run_in_terminal tool to execute pnpm test and file-specific test commands. This is the primary intended function of the skill and aligns with its metadata.
[PROMPT_INJECTION]: The skill presents an inherent surface for indirect prompt injection due to its requirement to read and process untrusted project source code.
Ingestion points: Source code and configuration files read via read_file (SKILL.md).
Boundary markers: None present. The agent is not explicitly instructed to ignore instructions found within the code it reads.
Capability inventory: Terminal execution (run_in_terminal) and file creation (create_file).
Sanitization: No sanitization or validation of the ingested content is specified.

test-engineer