Know what actually needs revalidation

A new model, agent platform, or protocol release can invalidate compatibility claims, pricing tables, output contracts, permission assumptions, and fixtures across a portfolio before the owner knows which products are exposed.

What it does

• Maps owner-supplied, source-linked API, pricing, reasoning, authorization, tool, protocol, safety, and output-contract changes only to products with matching declared exposure.
• Produces a deterministic CLEAR, REVIEW, or HOLD portfolio posture and a prioritized item-by-change retest queue.
• Separates documented release facts from inferred impact and names affected claims, evidence gaps, artifacts, and permissions.
• Preserves CANNOT_ASSESS when semantic work lacks both a compatible runner state and runner-contract evidence.
• Writes JSON and Markdown reports with a stable SHA-256 receipt for repeated comparisons.

Designed for honest modernization

A mapped impact is a retest obligation, not proof of a regression. The package runs offline, makes no provider call, edits no product, and never treats release attention, deployment, testing, owner verification, or marketplace activity as demand or revenue.

Included

The package includes the SKILL.md workflow, a standard-library Python mapper, input-contract reference, DeepSeek V4-Pro fixture, generated example reports, and seven behavioral and failure-case tests.