write-script-duckdb

Pass

Audited by Gen Agent Trust Hub on Aug 18, 2026

Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: Instructions guide the agent to perform script lifecycle tasks using the wmill CLI, such as testing local scripts with 'wmill script preview' or running deployed versions with 'wmill script run'.- [DATA_EXFILTRATION]: The skill documents methods for connecting to external databases using resources ('ATTACH $res:...') and writing query results directly to S3 storage buckets via the 'COPY TO' command.- [INDIRECT_PROMPT_INJECTION]: The skill defines a surface for processing untrusted data by reading CSV, Parquet, or JSON files from S3 storage.
  • Ingestion points: DuckDB reader functions and S3Object script parameters (SKILL.md).
  • Boundary markers: None identified in the provided templates.
  • Capability inventory: The skill facilitates S3 file access (read/write), external database connections, and script execution via CLI.
  • Sanitization: External data is processed without explicit sanitization instructions.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 18, 2026, 04:44 AM
Security Audit — agent-trust-hub — write-script-duckdb