The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes local shell commands and scripts to manage workflow state and configuration. Specifically, in SKILL.md, it runs python3 {project-root}/_bmad/scripts/resolve_customization.py to merge customization settings. In steps-e/step-e-01-discovery.md, it uses ls -t to find recent validation reports. These commands are used for workflow orchestration within the project's local directory structure.
[PROMPT_INJECTION]: The skill exhibits an indirect prompt injection attack surface (Category 8). It ingests untrusted data from multiple sources including user-provided PRD files, validation reports, and project context files (project-context.md).
Ingestion points: PRD markdown files, validation-report-*.md files, and {project-root}/**/project-context.md globs.
Boundary markers: The instructions do not specify explicit delimiters (like XML tags or unique markers) to isolate ingested file content from the agent's instructions.
Capability inventory: The skill possesses the ability to execute shell commands (ls), run Python scripts (python3), and write/update files on the local filesystem (PRD edits).
Sanitization: There is no evidence of sanitization or validation of the content read from external files before it is processed by the agent.

bmad-edit-prd