datahub-enrich

Pass

Audited by Gen Agent Trust Hub on Sep 17, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONDYNAMIC_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied metadata (descriptions, tags, glossary terms) which are written to an external DataHub catalog.
  • Ingestion points: User metadata input fields defined in SKILL.md.
  • Boundary markers: Includes instructions to ignore embedded commands but lacks structural delimiters in the target system.
  • Capability inventory: Bash execution of datahub CLI for metadata mutations.
  • Sanitization: Employs alphanumeric filtering and pattern stripping for metadata values.
  • [COMMAND_EXECUTION]: Executes the datahub CLI tool via Bash. The skill provides instructions to validate user input against shell metacharacters before passing it to the command line.
  • [DYNAMIC_EXECUTION]: Assembles temporary JSON variable files (/tmp/enrich-vars.json) using shell redirection to pass user-provided metadata to the datahub CLI.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 17, 2026, 09:44 AM
Security Audit — agent-trust-hub — datahub-enrich