The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes local Python scripts at activation and completion using the uv run command. These scripts (resolve_customization.py) are located within the project's _bmad/scripts/ directory, which is consistent with the vendor's development environment. These executions are triggered automatically during the workflow lifecycle.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from several sources: the data attribute at startup and external files referenced via globs in the persistent_facts configuration.
Ingestion points: SKILL.md Step 1 (data attribute) and Step 3 (persistent_facts file globs).
Boundary markers: The instructions do not define specific delimiters or "ignore instructions" warnings for these external inputs.
Capability inventory: The agent has the ability to execute subprocesses via uv run and write files to the project directory as defined in the default_output_file path.
Sanitization: There is no mention of sanitization or validation of the contents of the ingested files before they are treated as "foundational context."

bmad-cis-problem-solving