The Agent Skills Directory

[SAFE]: The skill operates as a legitimate creative assistant for film production. It correctly identifies the Renku project context and uses authorized tools for all operations, following established security best practices for AI agent skills.
[COMMAND_EXECUTION]: The skill invokes the 'renku' CLI to perform project-related tasks such as resolving scene IDs, listing shot plans, and importing media assets. These operations are restricted to the local environment and are documented with specific parameters that minimize risk.
[INDIRECT_PROMPT_INJECTION]: The skill manages an attack surface through the ingestion of scene and beat context from the project files. 1. Ingestion points: Retrieval of beat sheets and shot plan reports via 'renku screenplay' and 'renku shot-plan' commands. 2. Boundary markers: Shot descriptions are structured using a set of fixed Markdown headers (e.g., ## Intent, ## Camera & Optics) as defined in the writing guidelines. 3. Capability inventory: Modification of the shot plan state and asset selection via CLI commands. 4. Sanitization: The skill implements a mandatory validation step using 'renku shot-plan validate' before any create or update operation is executed.

shot-planner