The Agent Skills Directory

[SAFE]: The skill defines a logical workflow for file organization and metadata updates within a local workspace. It does not utilize external network connections, shell commands, or third-party dependencies.
[PROMPT_INJECTION]: The skill defines a process for ingesting external content which creates a surface for indirect prompt injection. 1. Ingestion points: Raw content processed via triggers like '讀這個' (Read this), '整理這份逐字稿' (Organize this transcript), or '幫我看這個 PDF' (Help me look at this PDF). 2. Boundary markers: No delimiters or instructions to ignore embedded commands are specified in the ingest workflow. 3. Capability inventory: The skill uses the agent's file system tools to read from Resources/ and write to memorbs/, Areas/, and Projects/. 4. Sanitization: No sanitization or filtering is prescribed for the external raw content before it is summarized or indexed.

memorb-ingest