data-auditor-cleaner
Pass
Audited by Gen Agent Trust Hub on Sep 3, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill operates entirely within the localized environment, specifically targeting paths within
workspace/data/andworkspace/code/. It enforces a read-only policy for raw data to ensure integrity. - [DATA_EXPOSURE]: The skill accesses local files (CSV, Excel, JSON) to perform its auditing functions. It does not contain any network-based tools (curl, wget) or attempts to access sensitive system directories such as SSH keys or environment variables.
- [DYNAMIC_EXECUTION]: The skill is designed to generate data cleaning scripts (e.g.,
data_cleaning.py). This code generation is a standard function of a data auditor and is governed by explicit rules for data transformation rather than arbitrary command execution. - [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from the
data_rawdirectory. While this represents a potential attack surface for indirect prompt injection, the workflow includes specific guardrails such as reading only column headers and samples first, and requiring user confirmation for mapping mappings.
Audit Metadata