The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes local Python scripts (resolve_customization.py, resolve_config.py) using uv run during its activation and completion phases. It also contains logic to execute arbitrary instructions provided via the on_complete, activation_steps_prepend, and activation_steps_append configuration fields.
[EXTERNAL_DOWNLOADS]: The skill suggests the installation of the core bmad skill from the bmad-code-org GitHub organization using the npx skills add command if the environment is not already configured.
[INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests and synthesizes data from untrusted sources, including local documents and web search results.
Ingestion points: The Artifact Analyzer agent scans local project directories for documents, and the Web Researcher agent retrieves content from the internet.
Boundary markers: No explicit delimiters or instructions to ignore embedded commands are used when processing the ingested data.
Capability inventory: The skill has access to the local file system, web search tools, and subprocess execution capabilities via uv run.
Sanitization: There is no evidence of sanitization or filtering of the external content before it is processed by the agent.

bmad-prfaq