update-graft-inventory

Pass

Audited by Gen Agent Trust Hub on Feb 20, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [Command Execution] (LOW): The skill utilizes npx wrangler d1 execute --remote to perform read and write operations on a production database. This provides high-privilege access to production infrastructure as part of its documented workflow.\n- [External Downloads] (SAFE): Uses npx to run the wrangler CLI tool. As wrangler is a standard tool maintained by Cloudflare, it is considered a trusted dependency for this environment.\n- [Indirect Prompt Injection] (LOW): The skill processes untrusted content from migration files and database query results, creating a surface for indirect prompt injection during metadata extraction.\n
  • Ingestion points: SQL migration files (libs/engine/migrations/*.sql) and database query outputs.\n
  • Boundary markers: Absent. The instructions do not define delimiters to separate untrusted data from processing logic.\n
  • Capability inventory: Remote database modification (wrangler d1 execute), file system writes, and repository commits.\n
  • Sanitization: No explicit sanitization or validation of the SQL content or query results is performed before the data is used to update the inventory and type definitions.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 20, 2026, 11:13 PM