The Agent Skills Directory

[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection through the processing of book content.\n
Ingestion points: Files in Markdown, TXT, and EPUB formats are ingested from local directories by ingest_bridge.py and search.py.\n
Boundary markers: Snippets of text from books are returned to the agent in search.py without markers to delineate data from instructions or warnings to ignore embedded commands.\n
Capability inventory: The skill is authorized to use Bash and Python tools according to SKILL.md, providing an execution path for injected instructions.\n
Sanitization: No sanitization or filtering is applied to the extracted text before it is presented to the agent.

consume-book