The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes a local Python script located at {project-root}/_bmad/scripts/resolve_customization.py during its activation sequence. This execution relies on the integrity of the project environment and the provided script.
[COMMAND_EXECUTION]: The skill is designed to sequentially execute steps defined in activation_steps_prepend and activation_steps_append. These are loaded from customize.toml and can be overridden by user or team configuration files, allowing the execution of arbitrary instructions or commands stored in external files.
[PROMPT_INJECTION]: The skill possesses a significant surface for indirect prompt injection. Ingestion points: The skill reads local files such as project-context.md and config.yaml, and it ingests data from numerous web searches conducted throughout the workflow (Steps 2 through 6). Boundary markers: No delimiters or defensive instructions are implemented to prevent the agent from obeying instructions embedded in the ingested data. Capability inventory: The skill has the capability to execute shell commands via the Python script and activation steps, and it can write research reports to the local filesystem. Sanitization: No sanitization or validation of the content retrieved from the web or project files is performed before it is processed or written to the final output documents.

bmad-technical-research