The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The orchestrator script (scripts/reverse-outline.ts) fetches utility functions from the official Deno registry via URL imports. This is standard Deno behavior and targets a well-known service.\n- [COMMAND_EXECUTION]: The skill uses an orchestrator script (scripts/reverse-outline.ts) to execute sub-processes for each stage of the analysis pipeline, which is a functional requirement for its automated workflow.\n- [DYNAMIC_EXECUTION]: The analysis workflow involves the orchestrator script dynamically identifying and running local scripts within the skill's directory to process the book data.\n- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted text from user-provided book files, which inherently exposes an analysis surface for indirect prompt injection if the source text contains adversarial patterns. As this is associated with the core evaluative purpose, the risk is minimal and characteristic of text analysis tools.\n
Ingestion points: Raw text from book.txt is ingested by segment-book.ts, analyze-scene-batch.ts, detect-genre.ts, and track-characters.ts.\n
Boundary markers: No explicit markers are used to separate book content from agent instructions during processing.\n
Capability inventory: File system read/write and sub-process execution capabilities are used across reverse-outline.ts, segment-book.ts, and the other analysis scripts.\n
Sanitization: The input text is analyzed using regular expression patterns without prior sanitization or filtering of potential injection strings.

reverse-outliner