The Agent Skills Directory

[PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface (Category 8) as it is designed to ingest and process external project materials such as repository files, PRDs, and website content.
Ingestion points: The bp-outline module reads project files and external URLs to generate investor narratives.
Boundary markers: The skill does not implement specific delimiters or 'ignore' instructions to isolate untrusted data within its prompt templates.
Capability inventory: The toolchain can write to the local filesystem, execute Python auditing scripts, and invoke the external lovstudio-any2deck skill.
Sanitization: While path validation is present in init_bp.py to prevent directory traversal, there is no explicit sanitization for the content of ingested documents.
[COMMAND_EXECUTION]: The skill utilizes local Python scripts (scripts/init_bp.py and scripts/audit_bp.py) to initialize workspaces and perform content linting. These scripts include security controls that prevent the skill from operating in sensitive directories like the system root or the user's home directory root.
[EXTERNAL_DOWNLOADS]: The skill kit references dependencies such as lovstudio-any2deck and the lovstudio CLI. These resources are recognized as originating from the verified vendor (lovstudio) and are consistent with the skill's primary purpose.

lovstudio-bp