recipe-post-mortem-setup

Pass

Audited by Gen Agent Trust Hub on Apr 1, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes the 'gws' command-line interface to manage Google Docs, Calendar events, and Chat messages. This is legitimate functionality for a Google Workspace automation skill.\n- [PROMPT_INJECTION]: The recipe includes placeholders such as '[Incident]' that are intended to be filled with external data. This creates a surface for indirect prompt injection if the input content is sourced from untrusted incident reports.\n
  • Ingestion points: Incident placeholders in the steps defined in SKILL.md.\n
  • Boundary markers: None present in the recipe instructions.\n
  • Capability inventory: Writing to Google Docs, inserting Calendar events, and sending Chat notifications via the 'gws' tool.\n
  • Sanitization: No explicit validation or sanitization of the input data is described in the recipe.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 1, 2026, 02:24 AM