Skip to content

Candidate v0.5.0-rc.1

v0.5.0-rc.1 starts Sprint 3—Practice and evidence. It preserves the stable v0.4.0 core and adds the first instrumented Practice Lab case.

Included

  • verified catalog of Claude and ChatGPT Work portfolios;
  • reasoned selection of the B2B Proposal case;
  • reproducible baseline using simulated demo data;
  • PDAMR contract, alternatives and decision rights;
  • YAML record, CSV log and process-quality rubric;
  • hold-out and difference-in-differences protocol;
  • ADR-018 and ADR-019;
  • gates against premature causal claims.

Evidence status

The case is instrumented-not-executed. There is no confirmed real owner, chosen option, preregistered expectation, action, outcome or attributable effect. The release demonstrates specification and traceability, not effectiveness.

Critical finding

Event and referral show higher observed win rates than digital inbound, but their confidence intervals overlap. The initial wording “significantly higher” is rejected. Budget reallocation remains blocked until evidence improves. This is cycle 0 of editorial learning: the evidence contract changed a recommendation. It is not evidence of commercial impact.

Unchanged

The v0.4.0 core authority, PULSE constitutional sources, CDI's proposed-domain status and absence of a formal WCAG claim all remain unchanged.

Publishing status

The candidate was deployed and documented through PR #6. Editorial, technical, visual and Axe gates passed. It has not been promoted to stable and remains untagged under ADR-022. Organizational data, real-owner confirmation and execution remain necessary to produce impact evidence.