The Agent Skills Directory

[COMMAND_EXECUTION]: The scripts/check.sh entry point executes a local Python runner (ds-dq.py) and project-specific test or lint commands defined within the data science plan. These commands are restricted to the local environment and intended for project validation.
[DYNAMIC_CONTEXT_INJECTION]: SKILL.md contains dynamic context placeholders using the !command`` syntax. These are used to execute utility scripts like skill-toc and load-constraints from the platform's plugin root to dynamically populate the skill documentation with a Table of Contents and constraint index at load time.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data during the profiling and planning phases, which is used to populate a ## Data Outputs table. This table is subsequently parsed by scripts/ds-dq.py. The skill implements structural boundaries by requiring a human-review phase and using a specific ds-reviewer agent to judge findings.
[DYNAMIC_EXECUTION]: The test suite scripts/ds_dq_test.py uses the subprocess module to execute the ds-dq.py script against fixtures to verify that data quality checks correctly identify failures and passes.

ds