The Agent Skills Directory

[COMMAND_EXECUTION]: The troubleshooting section instructs the agent to execute local scripts within the workspace using commands like uv run python scripts/pipeline.py kickoff|init and uv run python scripts/validate_repo.py.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process data from external templates and pipeline definition files (pipelines/*.pipeline.md, templates/UNITS.*.csv, and PIPELINE.lock.md) which could contain malicious instructions designed to influence agent behavior during the unit generation process.
Ingestion points: The skill reads PIPELINE.lock.md, pipelines/*.pipeline.md, and templates/UNITS.*.csv during the procedure (SKILL.md).
Boundary markers: None are specified to delimit template content from system instructions or to warn the agent against executing instructions found within the data.
Capability inventory: The skill has the capability to write to the local filesystem (UNITS.csv, STATUS.md, CHECKPOINTS.md) and execute shell commands/scripts in the scripts/ directory.
Sanitization: There are no explicit instructions for the agent to sanitize or validate the content of the pipeline or template files before processing them.

unit-planner