GROW-S2 | Systems | Free & open

Evaluation Auditability

Build a defensible evaluation and audit layer for agents in science, civic systems, and other high-stakes workflows.

8 skill artifacts · Draft · GROW-S2 (Systems cluster)

What this module contains

8 reusable skill artifacts.

Each artifact is a standalone, versioned build deliverable — templates, registers, policies, specs, checklists, matrices, playbooks, and maps — wired to the rest of the GROW stack by explicit contracts.

Eval Test Seed Matrix

matrix · v0.2.1

The Eval Test Seed Matrix is the official C1 consumer artifact named in substrate/interface-contracts.md.

Builds on: s1-failure-mode-register

Evaluation Charter

policy · v0.2.1

The Evaluation Charter declares what a GROW-built system must prove before it is allowed to act on behalf of a user, a constituent, or an operating business.

Builds on: s1-failure-mode-register

Audit Trail Schema

schema · v0.2.2

This schema is the contract with GROW-S3 (Data Provenance).

Contract: Produces C3 → Provenance

Builds on: c9-execution-accountability, e6-routing-policy, s2-evaluation-charter

Evaluator Roster

matrix · v0.2.0

The roster names who — human or automated — is allowed to sign off on what, when they are invoked, the scope of their authority, what artifacts their sign-off produces, and where escalation goes when they cannot or should not decide alone.

Builds on: s2-evaluation-charter, s1-failure-mode-register

Test Architecture

spec · v0.2.2

A test architecture is conformant when:

Builds on: s1-adversarial-test-plan, s1-failure-mode-register, s2-eval-test-seed-matrix, s2-evaluation-charter, s2-evaluator-roster

Scoring System

spec · v0.2.2

The Scoring System turns raw test verdicts and provenance records into a single, defensible scorecard with explicit pass thresholds, severity weighting, confidence handling, and drift indicators.

Builds on: s1-failure-mode-register, s2-audit-trail-schema, s2-eval-test-seed-matrix, s2-evaluation-charter, s2-evaluator-roster, s2-test-architecture, s3-decision-traceability, s3-provenance-metadata-schema, s3-provenance-package-templates, s3-source-inventory-template

Regression Discipline

playbook · v0.2.1

Regression discipline is how GROW prevents the most common operational failure: a "small change" that quietly breaks a previously-passing behavior.

Builds on: s2-evaluation-charter, s2-test-architecture, s2-audit-trail-schema, s2-scoring-system, s2-evaluator-roster, s3-reproducibility-controls

Audit Package Templates

template · v0.2.0

Six fillable scaffolds.

Builds on: s2-evaluation-charter, s2-test-architecture, s2-audit-trail-schema, s2-evaluator-roster, s2-scoring-system, s2-regression-discipline

Build & gate status

Where this module stands.

Status

Systems cluster — validated by the v2 integration gate (READY WITH CORRECTIONS).

How it composes

This module plugs into the four-cluster GROW stack (Systems → Execution → Trust → Coordination) through versioned interface contracts C1–C12. See the full wiring in the library registry.

Open access

GROW-S2 is free and open.

Every artifact, worksheet, and template in this module is free to use — no membership, no paywall.