The Agent Skills Directory

[COMMAND_EXECUTION]: The skill invokes local Deno scripts (scripts/representation-map.ts and scripts/sensitivity-audit.ts) to perform text and data analysis. These scripts are executed with the --allow-read flag, restricting their capabilities to file system reading only.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze untrusted content from user manuscripts. While the content is analyzed as static text and not executed, it represents an attack surface for indirect prompt injection.
- Ingestion points: scripts/sensitivity-audit.ts reads manuscript text files; scripts/representation-map.ts reads character JSON files.
- Boundary markers: Absent; the scripts read the entire target file content into memory for processing.
- Capability inventory: The skill can read local files that the Deno process has permissions for and write analysis results to the local filesystem.
- Sanitization: No explicit sanitization of manuscript content is performed, though the data is handled strictly through regular expressions and standard JSON parsing.

sensitivity-check