deck-refresh

Pass

Audited by Gen Agent Trust Hub on May 4, 2026

Risk Level: SAFE
Full Analysis
  • [Processing of External Content]: The skill is designed to ingest and process data from external sources, including user-provided Excel files, pasted text, and existing presentation slides.
  • Evidence: Phase 1 and Phase 2 describe reading slides, uploaded Excel sheets, and user-provided mappings to identify numbers for replacement.
  • Context: Reading external content is necessary for the skill's purpose. However, such content can sometimes contain instructions that might unexpectedly influence an agent's logic (indirect prompt injection). The skill mitigates this risk by requiring the agent to present a clear 'change plan' to the user and obtaining explicit approval before any modifications are executed.
  • [Document Modification Capabilities]: The skill possesses the ability to modify live presentations or regenerate slide files with new data.
  • Evidence: Phase 4 describes editing text runs, table cells, and chart data directly or writing results back to files.
  • Context: This is the core functionality of the skill. The instructions emphasize making the smallest possible changes and preserving existing formatting, which helps maintain the integrity of the original document. The multi-phase process with an 'approval gate' ensures that the user remains in control of the automated editing process.
Audit Metadata
Risk Level
SAFE
Analyzed
May 4, 2026, 09:17 PM