The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process data from external sources, including PowerPoint presentations and Excel spreadsheets. This creates a surface for indirect prompt injection where malicious instructions could be embedded in the data. However, the skill implements a robust mitigation strategy by requiring a mandatory human approval gate (Phase 3) where the agent must present a detailed plan of all changes before execution. This ensures that any unexpected or malicious behavior triggered by the input data is visible to the user before the skill exercises its file-writing capabilities.

deck-refresh