What problem does it solve? Operating a validated GxP laboratory information management system like LabWare LIMS carries serious compliance risk: authorizing a result is a 21 CFR Part 11 e-signature, an OOS result can be wrongly invalidated, and a batch can be released over an open failure. This Skill gives an AI agent the operator judgment to classify every LabWare action as read, reversible write, committing write, or destructive, and to stop consequential ones for human approval. ## Core Features & Use Cases - Risk-classified operation matrix: Maps sample login, result entry, review/authorization, batch disposition, COA issuance, and spec edits into read / reversible / committing / destructive classes with gating rules. - OOS/OOT and stability guidance: Covers the laboratory-phase OOS investigation, retest vs resample rules, stability study pulls, and the seams to the QMS (deviations, CAPA) and ERP (stock disposition). - Guardrails and recovery paths: Enforces segregation of duties, immutable audit-trail correct-forward patterns, and refusal rules for spec-gaming, testing into compliance, and releasing over an open OOS. - Use Case: An analyst asks the agent to release a batch while a dissolution result is flagged OOS; the Skill classifies the release as destructive, blocks it, and routes the work through the OOS investigation playbook instead. ## Quick Start Ask the agent to walk through entering, reviewing, and authorizing an assay result in LabWare and releasing the batch on spec.
