Governance Log

All architectural decisions GOV-001 through GOV-015.

Governance Log

Writing Production Studio — ProActive ReSolutions / Richard

This log records all deliberate platform decisions — changes to patterns, modules, architectural principles, and platform scope. The purpose is to make the studio's evolution legible over time and to prevent unintentional drift from documented decisions.

Governance Cycle

Cadence: Quarterly review of: pattern performance metrics, knowledge base quality indicators, outstanding retrospective findings, and any pending changes in the maintenance queue.

Annual: Voice profile review and update.

Triggered: When a pattern retrospective flags a High priority change; when a submission outcome reveals a significant intelligence gap; when a new output type is added to scope.

Log Format

[GOV-ID] — [Date] — [Type: Architecture / Pattern / Module / Scope / Process]

Decision: [What was decided] Rationale: [Why] Affected artifacts: [Which files were changed] Initiated by: [Retrospective / Submission outcome / Research finding / Richard's judgment]

Log

GOV-001 — March 2026 — Architecture

Decision: Platform instantiated. Seven-layer architecture adopted. Four-mode HITL model adopted. WSD schema adopted (revised structure per Research Findings v1.0, placing Locked Commitments at top and Session Brief at bottom). Rationale: Initial platform design based on broad research sweep and targeted thread research. See Writing_Production_Studio_Memo.md and Writing_Production_Studio_Research_Findings.md. Affected artifacts: All initial scaffold files. Initiated by: Design session, March 2026.

GOV-002 — 9 March 2026 — Scope / Architecture

Decision: Build sequence re-ordered. Practitioner/thought leadership content production (Phase 1A) takes priority over academic patterns (Phase 1B). Professional Voice Profile (Section B) built before Scholarly Voice (Section A). Practitioner patterns (PATTERN-PRACTITIONER-ARTICLE, PATTERN-TRADE-ARTICLE, PATTERN-OPINIONCOMMENTARY) built before PATTERN-BOOK-CHAPTER. The two tracks run in parallel after initial practitioner infrastructure is in place. Rationale: ProActive ReSolutions' immediate business need is Canadian brand re-establishment and sales growth. The original build sequence optimized for academic credibility (book chapter first, practitioner content at ~12 weeks). This means the Studio would produce no business-relevant output for months. Re-sequencing puts first published content within 2 weeks while preserving the academic track on a parallel timeline. The Studio's architecture (WSD, HITL modes, governance) applies equally to both tracks — no architectural changes required, only build order. Affected artifacts: BUILD_LOG.md (re-sequenced), VOICE_PROFILE.md (Section B prioritized), new patterns to be created. Initiated by: Strategic alignment session with Richard, 9 March 2026. Context: ProActive is a 25-year international conflict transformation consultancy (32 countries, clients include Google, FBI). Current pain point is Canadian market re-establishment. Content pipeline serves marketing, sales, product development, and thought leadership simultaneously. Suzanne Stewart (VP Client Services) confirmed as content collaborator. Target referral channels: employment/labour lawyers, HR, executives/boards, H&S. Target sectors: healthcare, universities, mining, forestry, engineering, First Nations, policing, fire & rescue, government (local, provincial, federal).

GOV-003 — 10 March 2026 — Pattern / Process / Scope

Decision: Three platform enhancements: (1) Publisher Disclosure Matrix updated with 6 material changes — SAGE 3-tier classification, Taylor & Francis expanded guidance + AI image ban, STM Association meta-standard, AID Framework structured disclosure taxonomy, Tri-Agency source-specific acknowledgment, Science/AAAS misconduct-level enforcement. (2) Slash command suite expanded from 6 to 9 — added /citation-audit (with APA/Chicago style options), /coherence-review (with configurable adversarial reader parameter), /project-status (WSD-stage-based dashboard); redesigned /gap-audit around WSD stages instead of argument-map-only; standardized flag system across all commands (7 flags documented in AGENTS.md). (3) Stage 5.5 (Atomization & Distribution) added to PATTERN-PRACTITIONER-ARTICLE-v1 — hub-and-spoke model producing executive summary, LinkedIn insight posts, framework visual spec, case study excerpt from each anchor piece. Distribution Plan section added to WSD template. Rationale: Extracted from analysis of publication-institutional-listening project infrastructure. Publisher landscape has materially shifted since initial matrix was built. Slash commands generalized from argument-map-specific to WSD-stage-generic to serve all studio patterns. Stage 5.5 addresses the gap between publication and distribution — ProActive's brand re-establishment strategy requires content atomization across channels. Affected artifacts: PUBLISHER_DISCLOSURE_MATRIX.md, PATTERN-PRACTITIONER-ARTICLE-v1.md, WSD_TEMPLATE.md, AGENTS.md, all .claude/commands/ files. Initiated by: Handoff from publication-institutional-listening extraction analysis session, 10 March 2026.

GOV-004 — 11 March 2026 — Pattern

Decision: PATTERN-TRADE-ARTICLE-v1 created. Mid-weight pattern for trade/industry publication articles (1,500–3,000 words) with light sourcing requirements (5–12 references). Adds Stage 2 (Research & Source Assembly) and Stage 6 (Submission & Tracking) not present in PATTERN-PRACTITIONER-ARTICLE-v1. Integrates Confidence Schema practitioner adaptation, Verification Pipeline Layer 1 (mandatory) and Layer 2 (recommended), and audience-specific sourcing strategy. Stage 6.5 atomization extends Stage 5.5 with blog derivative spoke. Seven stages total: Topic/Angle/Venue → Research & Source Assembly → Structural Outline → Draft → Review & Revision → Submission & Tracking → Atomization & Distribution. Rationale: GOV-002 identified PATTERN-TRADE-ARTICLE as a priority build. PATTERN-PRACTITIONER-ARTICLE-v1 already cross-references it as the appropriate pattern for longer trade publications. Trade publications occupy a distinct niche: more substance and sourcing than blog/LinkedIn content, but lighter than academic patterns. The pattern bridges the practitioner and academic tracks. Affected artifacts: patterns/PATTERN-TRADE-ARTICLE-v1.md (new), AGENTS.md (updated build state and directory listing). Initiated by: Platform build session, 11 March 2026.

GOV-006 — 12 March 2026 — Architecture / Knowledge Base

Decision: Voice profile system redesigned. Single flat VOICE_PROFILE.md superseded by a multi-tier architecture: (1) _schema/ layer defining shared vocabulary (NN/G Four Dimensions + ghostwriting markers + Discourse Ground taxonomy); (2) speaker-first directory hierarchy (richard-hart/, proactive-resolutions/, clients/); (3) persistent voice-profile.md per speaker; (4) context-guides/ per speaker for contextual application; (5) calibration-log.yaml per speaker for piece-level calibration records. Key architectural concept: Discourse Ground — the prior and enabling context of a communication (brand, field/discipline, platform, or hybrid). Replaces the discarded "speaker integrity vs. listener-fit" bi-polar framing, which has no established basis in professional writing methodology. Rationale: Original voice profile design was a single flat document with no listener dimension, no multi-speaker support, and no calibration mechanism. External research (NN/G, CMI, ghostwriting methodology — research agent, 2026-03-12) showed: (1) speaker integrity vs. listener-fit is not a recognized framework — industry integrates them through calibration; (2) NN/G Four Dimensions is the established baseline schema; (3) ghostwriting practice adds a speaker-markers layer (vocabulary, rhythm, analogy patterns, rhetorical moves) that dimensional scales miss; (4) two-tier model (persistent profile + piece-level calibration) is supported by practitioner evidence. Dialogue refinement identified that the organizational level above speakers is better named Discourse Ground: the prior, enabling context of any communication — which can be organizational brand, personal brand, scholarly field, platform, or hybrid. Not all communications are brand-framed; scholarly and field-governed communications need the concept to work. Source registry (best-practice-sources.yaml) updated with writing/professional category (NN/G, CMI, Professional Ghost, RSA). Resolves L-177 source gap. Affected artifacts: knowledge-base/voice-profiles/ (new), knowledge-base/voice-profile/ (removed), best-practice-sources.yaml in dev-env-docs (writing/professional category added), _architecture/component-registry.yaml (updated). Initiated by: Design critique session with Richard, 12 March 2026. Research validated by agent web research.

GOV-007 — 13 March 2026 — Pattern / Architecture / Process

Decision: Platform enhancements in two phases (same session, GOV-007a and GOV-007b):

GOV-007a — Five enhancements from burnout v2 external review: (1) L-186, L-187, L-188 captured in global LESSONS-LOG. (2) PATTERN-STAGE3-REVIEW-PROTOCOL-v1 → v1.1: R10 (Originality/Positioning Reviewer) added, runs first in all sequences. (3) PATTERN-PRACTITIONER-ARTICLE-v1 Stage 1 enriched: Minimum Viable Discourse Scan (all content types, calibrated by type), Experience Intent, Conclusion Intent, Evidence Selection. (4) Audience persona library initiated: knowledge-base/audience-personas/healthcare-hr-director.md (VPC + JTBD + discourse landscape). (5) Burnout article v3 drafted: title changed ("Before You Commission Another Wellness Initiative"), opening restructured (situation-list, reader-addressed), compound exhaustion expanded (cognitive/relational/physiological), Stage 3 verdict: PROCEED TO STAGE 4.

GOV-007b — Further enhancements during QC remediation pass: (6) PATTERN-STAGE3-REVIEW-PROTOCOL-v1 → v1.2: R11 (Prose/Craft Editor) added — reads for standout moments, dead weight, phrasing originality, rhythm; max 7 surgical observations; runs last in sequence; mandate boundary defined. NOTE: R11 is NEEDS_UPDATE — external validation against craft editing methodology required by 2026-03-20 before first production use. (7) PATTERN-TRADE-ARTICLE-v1 Stage 1 enriched: preliminary discourse scan added, Experience/Conclusion/Evidence intents added; Stage 2 Discourse Positioning Section added (3-5 adjacent voices, echo risk); Stage 3 craft gate question added; bidirectional revision rules added (cross-references PRACTITIONER + 4 trade-specific rows). (8) Bidirectional Revision Rules added to PATTERN-PRACTITIONER-ARTICLE-v1 (decision table: when to return to Stage 1 vs. proceed inline). (9) templates/wsd/WSD_TEMPLATE.md Section 1 extended: Discourse Scan result, Experience Intent, Conclusion Intent, Evidence Selection fields added. (10) continuity-index.yaml schema block and outcome fields added; collection protocol documented. (11) Component Registry updated: PRACTITIONER, TRADE, STAGE3-PROTOCOL, WSD-TEMPLATE entries updated; healthcare-hr-director.md and burnout v3 registered. (12) Signal log seeded with 6 signals (2 with assessment_needed: true for R11 validation).

Rationale: External review (Originality_Analysis_and_Revision_Strategy.md, 2026-03-13) identified echo risk in burnout v2. This was a Stage 1 design failure with a Stage 3 detection gap. Both gaps closed. Discourse scan extended to ALL content types per user direction. R11 fills functional coverage gap (craft was unreviewed). TRADE pattern brought to parity with PRACTITIONER enrichments.

Affected artifacts: patterns/PATTERN-STAGE3-REVIEW-PROTOCOL-v1.md (v1.2), patterns/PATTERN-PRACTITIONER-ARTICLE-v1.md (Stage 1 + bidirectional rules), patterns/PATTERN-TRADE-ARTICLE-v1.md (Stage 1 + Stage 2 + bidirectional rules), knowledge-base/audience-personas/healthcare-hr-director.md (new), project-pipeline/drafts/burnout-misdiagnosed_AI_draft_v3.md (new), templates/wsd/WSD_TEMPLATE.md (Section 1 extended), knowledge-base/voice-profiles/richard-hart/continuity-index.yaml (schema + outcome fields), _architecture/component-registry.yaml (updated), _architecture/signal-log.yaml (seeded), ~/dev/infrastructure/dev-env-docs/LESSONS-LOG.md (L-186, L-187, L-188).

Initiated by: Burnout article v2 external review session, 13 March 2026. QC remediation pass completed in same session.

GOV-008 — 24 March 2026 — Architecture

Decision: Cascade system architecture adopted. Four-level hierarchy: Level 1 Per-Piece (WSD + Distribution Brief + Performance Record), Level 2 Sector/Audience Intelligence (Sector files + Personas + Referral Network), Level 3 Quarterly Planning (Quarterly Review + Q-Portfolio), Level 4 Annual Strategic (Business Strategy Brief + Content Strategy). Five named feedback loops with explicit triggers: Loop A (Performance→Strategy, quarterly), Loop B (Audience Signal→Personas, 30-day), Loop C (Sector Signal→Intelligence, monthly), Loop D (Referral Signal→Network, per-conversation), Loop E (Pattern Performance→Pattern Files, quarterly). Three-tier metrics framework: Tier 1 Reach (lagging), Tier 2 Engagement (EQS leading indicator), Tier 3 Business Impact (ultimate). Fifteen new artifacts created spanning all four levels. Rationale: Maps analysis (Ecosystem Map, Artifact Flow Map, Content Lifecycle Workflow) identified the Studio as "step 4 of 11 in the content value chain" with missing upstream (strategy) and downstream (distribution, measurement, learning) infrastructure. Cascade architecture fills the downstream gaps. Feedback loops close the system so performance data returns to influence content decisions. External research agent confirmed cascade architecture resonant with content marketing operations best practice. Affected artifacts: _architecture/CASCADE_SYSTEM_ARCHITECTURE.md (new — system blueprint), intelligence/BUSINESS_STRATEGY_BRIEF.md (new), intelligence/REFERRAL_NETWORK.md (new), intelligence/sectors/ (7 new sector intelligence files), knowledge-base/audience-personas/employment-labour-lawyer.md, ohs-professional.md, executive-board.md (3 new personas), templates/projects/DISTRIBUTION_BRIEF_TEMPLATE.md, PERFORMANCE_RECORD_TEMPLATE.md, templates/QUARTERLY_REVIEW_TEMPLATE.md (3 new templates), projects/POPUP-2026-01/DISTRIBUTION_BRIEF.md, PERFORMANCE_RECORD.md (2 populated project artifacts). Initiated by: Strategic buildout session, 24 March 2026. User direction: "plug the gaps in the system you're suggesting, design underlying artifacts, cascade information up, feedback loops, metrics, based on current best practice."

GOV-009 — 24 March 2026 — Process

Decision: Distribution Brief created before production begins. The cascade plan (primary publication, LinkedIn post sequence with draft opening lines, newsletter placement, speaking use, proposal language, repurposing triggers) is locked before any drafting starts. Prevents the "dead-end publication pattern" where content is produced but distribution is improvised or absent. Rationale: Content Strategy Q2 Distribution Plan (CONTENT_STRATEGY.md) established the pattern; GOV-008 cascade build formalized it as a per-project artifact. The Distribution Brief is Level 1 infrastructure — it must precede production, not follow it. Affected artifacts: templates/projects/DISTRIBUTION_BRIEF_TEMPLATE.md (new), projects/POPUP-2026-01/DISTRIBUTION_BRIEF.md (new — first instance). Initiated by: Cascade build session, 24 March 2026.

GOV-010 — 24 March 2026 — Architecture

Decision: Single Source of Truth (SSOT) implementation pattern established. (1) EQS formula (saves×5 + long_comments×2.5 + short_comments×1 + reactions×1) is canonical in knowledge-base/publishing-intelligence/PERFORMANCE_TRACKER.md. All other artifacts reference it; none redefine it. (2) Audience hierarchy (Priority 1–5) is canonical in intelligence/BUSINESS_STRATEGY_BRIEF.md. CONTENT_STRATEGY.md references it. (3) Implementation mechanism: include-markdown plugin for shared calculated formulas; cross-reference links for stable reference sections. Metadata headers in every artifact declare what each owns and what each references. Rationale: SSOT violations identified in the same session that introduced the principle (EQS formula copied to 3 locations; audience hierarchy in 2 locations). Research agent finding: "Every manual process becomes a source of drift in a small team." SSOT without enforcement produces drift from day one. Affected artifacts: PERFORMANCE_RECORD_TEMPLATE.md (EQS reference added; not redefined), templates/QUARTERLY_REVIEW_TEMPLATE.md (same), knowledge-base/content-strategy/CONTENT_STRATEGY.md (audience hierarchy canonical reference note added), all new artifacts (YAML frontmatter metadata headers added). Initiated by: Cascade build session, 24 March 2026. Research agent external validation.

GOV-011 — 24 March 2026 — Architecture

Decision: Audience Persona methodology standardized as VPC + JTBD: Strategyzer Value Proposition Canvas (Customer Jobs / Pains / Gains) plus Jobs-to-be-Done Operative Formula ("When I…, I want to…, so I can…, without…"). Discourse Ground section added per Voice Profile architecture (GOV-006): register, evidence type, tone, and what to avoid for each audience. Standard sections: Profile Summary, VPC, JTBD Operative Formula, Discourse Ground, Content Calibration Notes. Rationale: First persona (healthcare-hr-director.md, GOV-007) established VPC + JTBD as the pattern. Standardizing ensures all personas are built to the same methodology and are directly usable for content calibration decisions. Affected artifacts: knowledge-base/audience-personas/employment-labour-lawyer.md, ohs-professional.md, executive-board.md (all new, follow standard). Initiated by: Cascade build session, 24 March 2026.

GOV-012 — 24 March 2026 — Architecture / Process

Decision: Artifact metadata headers standardized. All artifacts carry YAML frontmatter: artifact_id, owns (SSOT declaration), references (canonical sources used), feeds (downstream artifacts), map_component (cascade level), owner, last_updated. Machine-readable by artifact monitor. Human-readable purpose: any new session can read an artifact's header and know its place in the system without reading CASCADE_SYSTEM_ARCHITECTURE.md. Rationale: Without metadata headers, the inter-artifact relationship graph exists only in the architecture document, not in the artifacts themselves. A new agent session starting from an artifact has no context for how it fits. Metadata headers make the system self-documenting. Affected artifacts: All 18 new artifacts from GOV-008 (headers added). Template for future artifacts: includes frontmatter block. Initiated by: Cascade build session, 24 March 2026.

GOV-013 — 24 March 2026 — Architecture / Process

Decision: Multi-signal artifact monitoring automation deployed. Monitors: (1) artifact count (WARN >25, CRIT >40), (2) SSOT violations — EQS formula or equivalent field definitions duplicated outside canonical source, (3) broken cross-references — relative links in markdown that resolve to missing files, (4) incomplete project artifact sets — projects missing WSD + Distribution Brief + Performance Record. Notification via ~/bin/notify.sh (ntfy + Telegram) on WARN or CRITICAL. Post-commit hook runs silently; output only on threshold breach. Architecture reassessment trigger: count >40 OR SSOT violations >3 → notification recommends AI-generated field registry and architecture review. Count-only threshold (original "25 artifacts" proposal) rejected as arbitrary and gameable; multi-signal threshold is more robust. Rationale: "25+ artifacts → change approach" stated without automation is governance theater. Automation-first principle (PRIN-001) applies. Single-metric thresholds fail when the metric can be gamed or misread. Multi-signal monitoring produces meaningful alerts, not noise. Affected artifacts: _architecture/tools/artifact-monitor.py (new), .git/hooks/post-commit (new), _architecture/tools/run-monitor.sh (new). Initiated by: Cascade build session, 24 March 2026. Research agent: "every manual process becomes a source of drift in a small team."

GOV-014 — 24 March 2026 — Architecture / Scope

Decision: Docs site (writing-studio.pages.dev) extended with two new sections: (1) Cascade System Architecture Map — 6th HTML map showing 4-level cascade, 5 feedback loops, 3-tier metrics, artifact locations. (2) Project Management section in MkDocs nav: artifact inventory page, active pipeline view, quarterly review process guide. Non-technical stakeholders (Suzanne, team reviewers) are the primary audience for the docs site. Every major architectural addition to the Studio must be reflected in the docs site. Rationale: 15 new artifacts and the cascade architecture were added without updating the docs site. Suzanne cannot navigate the system's relationships from the docs site without a cascade visualization. A "new architecture → new map" rule prevents documentation debt accumulation. Affected artifacts: docs/assets/maps/CASCADE_SYSTEM_ARCHITECTURE_MAP.html (new), docs/ project management pages (new), mkdocs.yml (nav extended). Initiated by: Cascade build session, 24 March 2026.

GOV-015 — 24 March 2026 — Process / Architecture

Decision: Three protocol gaps identified and logged; two require action before next session:

(a) Visual artifact pre-production checklist (OPEN — requires implementation): No pre-production checklist exists for visual artifacts (maps, diagrams, infographics). Signal-log 2026-03-18 (assessment_needed: true) flagged this after a FigJam service blueprint was produced with the wrong map type, wrong structural model, and wrong temporal assumptions. Required checklist items before producing any visual artifact: (1) What map type is this — structural, temporal, or process-flow? (2) What tool is fit for purpose for this type? (3) What are the quality criteria — who is the audience, what must they be able to read from it? (4) What governance artifacts does it need to align with? This checklist should be added to AGENTS.md under a "Visual Production" section or as a standalone _architecture/VISUAL_PRODUCTION_PROTOCOL.md.

(b) ARTIFACT_FLOW_MAP staleness (RESOLVED in this session): The map was not updated when Content Strategy and Sector Intelligence artifacts were created. Two MISSING entries corrected to present status. Root cause: no map update trigger exists when studio artifacts are created or modified. Recommended fix: add "Update ARTIFACT_FLOW_MAP.html if adding a new artifact type" as a mandatory step in GOV log template.

(c) AGENTS.md build state maintenance (RESOLVED in this session): AGENTS.md "Current Build State" section had a stale entry (Argument Bank "NOT YET BUILT" — superseded by ICP 2026-03-23). Corrected. Root cause: AGENTS.md is manually maintained with no update reminder. Recommended fix: add AGENTS.md build state check to the start of any session that creates or retires a major artifact.

Rationale: Runthrough session (2026-03-24) identified these gaps. Items (b) and (c) are resolved. Item (a) is open — implementation deferred to next visual production session.

Affected artifacts: _architecture/signal-log.yaml (3 new entries), AGENTS.md (Argument Bank line updated), _architecture/ARTIFACT_FLOW_MAP.html (2 MISSING entries corrected), projects/POPUP-2026-02/DISTRIBUTION_BRIEF.md (new), projects/POPUP-2026-02/PERFORMANCE_RECORD.md (new).

Initiated by: Studio runthrough session, 24 March 2026.

GOV-005 — 11 March 2026 — Process / Architecture

Decision: Continuous Excellence Protocol (PATT-024) infrastructure created. Component Registry (_architecture/component-registry.yaml) and Signal Log (_architecture/signal-log.yaml) initialized. 41 artifacts inventoried across all studio directories. First biweekly review scheduled for 2026-03-25. CLAUDE.md committed to the repository. Rationale: DEC-228 adopted PATT-024 for the Writing Studio as its founding adopter. The adoption gate requires a Component Registry to be in place. CLAUDE.md formalizes the operational rules. The signal log enables the event-driven review trigger mechanism. Affected artifacts: _architecture/component-registry.yaml (new), _architecture/signal-log.yaml (new), CLAUDE.md (committed to repo), _architecture/BUILD_LOG.md (updated). Initiated by: Platform build session, 11 March 2026. DEC-228 adoption context.