The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes local Deno scripts (scripts/init.ts, scripts/status.ts, scripts/watcher.ts) to automate project setup and file monitoring. These scripts require explicit file system permissions (--allow-read, --allow-write) provided by the user to operate on the story project directory. The logic is transparent and focused on recording file modification timestamps and paths for synchronization analysis.\n- [SAFE]: No network access or external data transmission was detected. The scripts do not contain any fetch calls or other network-related operations. No hardcoded credentials or sensitive system path access (outside the project scope) were identified.\n- [PROMPT_INJECTION]: The skill acts on untrusted data from the user's manuscript and story files. Ingestion points: The agent reads markdown files from the pitch/, structure/, scenes/, entities/, and manuscript/ directories. Capability inventory: The scripts are limited to file manipulation within the project path; no remote code execution or shell access is present. Sanitization: No specific delimiters or sanitization steps are defined for processing manuscript content, representing a surface for indirect prompt injection, though the risk is negligible given the localized nature of the skill and the absence of high-risk capabilities.

story-zoom