The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes a local Python script during activation: python3 {project-root}/_bmad/scripts/resolve_customization.py --skill {skill-root} --key agent. This command is used to merge configuration data from various project-level TOML files.
[REMOTE_CODE_EXECUTION]: The skill is designed to process and execute sequences of actions defined in the activation_steps_prepend and activation_steps_append fields within configuration files. This allows for dynamic execution of logic defined outside the core instruction set.
[PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection by ingesting external data into the agent's working context.
Ingestion points: Content is read from local files like project-context.md (via glob patterns), config.yaml, and multiple customization TOML files.
Boundary markers: There are no explicit delimiters or instructions provided to the agent to ignore potentially malicious directions embedded within these ingested files.
Capability inventory: The skill can execute shell commands through Python and perform multi-step automated sequences during persona activation.
Sanitization: No content validation or escaping is applied to the data retrieved from external files before it is integrated into the session context.

bmad-cis-agent-brainstorming-coach