The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to use bash utilities such as ls, head, sed, and grep to interact with the local file system. As documented in rules/corpus-guide.md, these commands are intended to list contents and extract specific text segments from a literary corpus stored in absolute paths like /Users/fred/Documents/重点文献/.
[INDIRECT_PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection because it ingests and processes large volumes of external data from a local corpus of approximately 800 works.
Ingestion points: Local text files within the /Users/fred/Documents/重点文献/ directory, accessed via shell commands.
Boundary markers: The instructions do not specify explicit delimiters or warnings to the agent to disregard any instructions that might be embedded within the literary texts.
Capability inventory: The agent is authorized to use shell commands for file reading (ls, head, sed, grep) and to perform extensive text generation and manipulation based on the ingested data.
Sanitization: The skill does not provide mechanisms for sanitizing or escaping the content retrieved from the corpus files before it is interpolated into the agent's context.

literary-master