The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the uv tool to execute a Python script located at {project-root}/_bmad/scripts/resolve_customization.py during its activation process. This script is a vendor-provided tool for managing configuration overrides within the project environment.- [DYNAMIC_EXECUTION]: The skill defines activation_steps_prepend and activation_steps_append fields in SKILL.md and customize.toml. The instruction logic in SKILL.md explicitly executes every entry in these arrays in order. While these arrays are empty in the default configuration, the implementation provides a mechanism for executing arbitrary commands or instructions defined in customization files.- [INDIRECT_PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection through its extensive data ingestion logic.
Ingestion points: The skill reads data from customize.toml, team overrides in {project-root}/_bmad/custom/{skill-name}.toml, user overrides in .user.toml, and a global config.yaml. Additionally, Step 4 in SKILL.md allows loading arbitrary file content based on glob patterns defined in persistent_facts.
Boundary markers: The skill lacks explicit delimiters or instructions to ignore instructions embedded within the loaded files or facts.
Capability inventory: The skill possesses the capability to execute commands (via activation steps) and dispatch intents to other skills or execute specific prompts.
Sanitization: There is no evidence of sanitization, validation, or filtering of the content loaded from these external files before it is interpolated into the agent's foundational context.

bmad-cis-agent-creative-problem-solver