reviewing-migration-and-data-safety
Pass
Audited by Gen Agent Trust Hub on Jul 6, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill's primary function is to provide an AI agent with a checklist for reviewing code diffs related to database persistence. The instructions in
SKILL.mdandreference/heuristics.mdalign with well-known database engineering patterns like the 'expand/contract' migration strategy. - [SAFE]: External references in
reference/sources.mdandreference/tool-rules.mdpoint to authoritative industry sources (e.g., Martin Fowler, Martin Kleppmann) and well-known open-source tools (e.g., GitHub's gh-ost, ankane/strong_migrations, and Percona's pt-online-schema-change). These references are provided for context and guidance, not for execution. - [SAFE]: Analysis of the evaluation scenarios in
evals/eval.jsonand code examples inexamples.mdshows they are strictly focused on identifying common migration risks (such as write-blocking table locks or unbatched updates) and do not contain any malicious payloads or bypass attempts. - [SAFE]: There are no indicators of data exfiltration, credential harvesting, or unauthorized command execution. The skill does not perform network operations or access sensitive local files.
Audit Metadata