spreadsheet-logic-builder
Pass
Audited by Gen Agent Trust Hub on Oct 4, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process user-provided spreadsheet definitions and samples to generate executable automation logic.
- Ingestion points: According to
SKILL.mdandreferences/dialogue-playbook.md, the skill prompts users to provide sheet names, column headers, and sample rows for matching and transformation logic. - Boundary markers: The workflow in
SKILL.mdincludes mandatory verification steps such as checking control totals and reviewing output tabs before final use, which serves as a manual oversight mechanism. - Capability inventory: The skill generates complex logic including Python scripts, VBA macros, and Excel formulas which are intended for local execution by the user as described in the engine selection section of
SKILL.md. - Sanitization: The skill explicitly includes a "Privacy rules" section in
SKILL.mdthat advises against uploading sensitive information (SSNs, EINs, bank accounts) and suggests using fake examples to define logic.
Audit Metadata