Skip to main content

Vision Alignment Synthesis — Cross-Perspective Recommendations

Date: 2026-02-17 Input: 4 perspective analyses (CISO, Product Owner, Developer, Architect) evaluating docs 06/07 against Sergey's W1 vision and UX.


Executive Summary

All four perspectives converge on a clear conclusion: docs 06/07's 4-concept separation is architecturally correct, but W1's product vision requires a fundamentally different user-facing approach — authority governance language, cluster-first navigation, and simplified deterministic paths instead of topology/chain concepts.

The naming questions in doc 07 (Q1-Q5) are largely deprioritized by W1 because W1's UX doesn't expose execution chains to users at all. The urgent work is building the W1-specific views, not renaming existing ones.


1. Consensus Naming Decisions (Q1-Q5)

Q1: What to call Concept 2 ("Execution Chains")?

PerspectiveRecommendation
CISOAuthority Chains — anchors governance narrative
Product OwnerKeep internally; don't surface in W1 — chains page not in W1 flow
DeveloperKeep "Execution Chains" internally; don't prioritize rename — zero W1 user impact
ArchitectAuthority Chains — survives beyond W1, aligns with 9-entity model

Decision: "Authority Chains" as the canonical internal/documentation name. Not surfaced in W1 UX at all.

  • Rename in docs and glossary now (low effort)
  • Keep execution_chains as the collection name in code (no storage rename)
  • Do not add Authority Chains to W1 navigation

Q2: Column header for execution count?

PerspectiveRecommendation
CISOExecutions (30d) — matches W1 UX, communicates magnitude
Product OwnerExecutions (30d) — W1 UX already decided this
DeveloperExecutions (30d) — aligns AutomationsPage with W1 vocabulary
ArchitectObserved Runs (30d) — more precise, survives into W2

Decision: "Executions (30d)" — 3-of-4 consensus, matches W1 UX spec exactly. The precision benefit of "Observed Runs" doesn't outweigh consistency with the W1 vocabulary the CISO will see.

Q3: Timestamp label on chain detail pages?

PerspectiveRecommendation
CISOLast Computed — prevents "chain ran at T" misreading
Product OwnerLast Refreshed — W1 vocabulary consistency
DeveloperLast Computed — most accurate for chain assembly
ArchitectLast Computed — unambiguous about sync-time derivation

Decision: "Last Computed" — 3-of-4 consensus. Chain detail is not a W1 page, so W1 vocabulary consistency matters less than semantic accuracy. W1 pages use "Last Refreshed" for their own global timestamp.

Q4: Glossary 3-concept vs 4-concept model?

PerspectiveRecommendation
All four(B) Add 4th concept now as "planned/W2"

Decision: Unanimous. Add AutomationRun to glossary as "planned — W2".

Q5: Nav item for chains page?

PerspectiveRecommendation
CISO"Authority Chains" + Shield icon, secondary to W1 flow
Product OwnerDon't include in W1 navigation at all
DeveloperKeep "Chains" but move to a collapsed developer/debug section
Architect"Authority Chains" + Shield, W1 excludes it from primary nav

Decision: Remove from W1 primary navigation. If retained for debug/power-user mode, label "Authority Chains" with Shield icon in a collapsed section.


2. Critical Gaps All Perspectives Identified

These are things Sergey's W1 vision demands that neither docs 06/07 nor the current platform address:

2.1 Risk Clusters (unanimous — highest priority gap)

All four analysts identified "Autonomous Authority Risk Clusters" as the biggest missing concept. W1 UX makes these the primary homepage navigation surface.

What's needed:

  • A deterministic aggregation pattern over active W1 findings
  • Compound condition definitions (e.g., "Sensitive + LLM + Active + Invalid Owner")
  • Fixed priority ordering
  • Server-side API endpoint (GET /api/v1/w1/risk-clusters)
  • New ADR documenting the concept (Architect recommendation)

Architecture: Computed view, not persisted collection. Aggregation over findings.

2.2 Standing Authority (CISO + Product Owner identified)

W1 UX features "Standing Authority" as a visually dominant section in every finding detail:

Execution Model: Autonomous
Authentication Type: Client Credentials
Human Session Required: No

Docs 06/07 never name or define this concept. It needs to be added as a named concept in the W1 model.

2.3 Execution Magnitude (CISO + Developer identified)

Doc 06 defines execution_status as binary (proven | unproven). W1 requires quantitative Executions (30d) alongside the binary status. The ExposureUnit shape needs execution_count_30d added.

2.4 New W1 Evaluator Rules (Architect + Developer identified)

Current evaluator has 4 rules. W1 requires at least 4 additional:

  • unknown_identity_binding — no deterministic RUNS_AS binding
  • reachable_sensitive_domain — identity can reach confidential/restricted data
  • llm_egress — connection path terminates at LLM endpoint
  • external_egress — connection path terminates at external endpoint

Plus refinements to existing rules for W1 scope.

2.5 Since Last Refresh Delta (CISO + Product Owner identified)

W1 UX requires "+X new autonomous identities, +Y ownership invalidations, -Z removed" delta display. No sync-to-sync comparison logic exists.


3. What Doc 07's Phase 1 Becomes Under W1

Doc 07 proposed 5 UI label changes as Phase 1. Here's the updated status:

Doc 07 Proposed ChangeW1 Status
Nav "Chains" renameMoot — W1 removes chains from primary nav
Page title "Execution Chains" renameMoot — page not in W1 flow
"Last Seen" → "Last Computed"Low priority — chain detail not in W1 flow
"Execution" → "Executions (30d)" on AutomationsPageYes, do this — good hygiene even if AutomationsPage isn't primary W1
Glossary update to 4-conceptYes, do this — add W2 placeholder

Net: 2 of 5 original changes remain relevant. The naming work shifts to building W1-specific surfaces.


4. W1 Implementation Plan (Cross-Perspective Consensus)

Phase W1-A: Foundation (Immediate, Lowest Risk)

#ItemEffortOwner Domain
1Update glossary to 4-concept + W2 placeholderSmallDocs
2Rename "Execution" column to "Executions (30d)" on AutomationsPageTrivialUI
3Fix stale API doc migration note (04-api-layer.md:140)TrivialDocs
4Fix chain-role naming mismatch (types.ts vs ADR-008)SmallBackend
5Define "Standing Authority" + "Authority Chain" in glossary and doc 06SmallDocs
6Restructure UI navigation: W1 primary (Homepage, Findings) + collapsed debug sectionSmallUI
7Implement computeW1ExposurePath() — bounded graph traversal for W1MediumBackend
8Implement posture summary endpoint (Active NHI, Dormant, Operator, Human counts)MediumBackend

Phase W1-B: Risk Clusters + Findings View (Core W1 UX)

#ItemEffortOwner Domain
9Write ADR for Risk Cluster conceptSmallDocs
10Implement risk cluster aggregation API endpointMediumBackend
11Build W1 Homepage (posture summary + delta + Top 5 risk clusters)LargeUI
12Build W1 Findings View (cluster-grouped, path row format, inline expand)LargeUI
13Finding enrichment endpoint (join finding to entity + path + evidence data)MediumBackend
14Add new W1 evaluator rules (unknown_identity_binding, reachable_sensitive_domain, llm_egress, external_egress)LargeBackend

Phase W1-C: Finding Detail + Polish

#ItemEffortOwner Domain
15Build W1 Finding Detail (standing authority block, 4-section evidence panel, linkage proof card)LargeUI
16Simplify path diagram to W1 format (4-node: Automation → Identity → Destination → Data Domain)SmallUI
17Since Last Refresh delta computation + displayMediumBackend + UI
18Sensitive domain summary displaySmallUI
19Remediation placeholder buttonTrivialUI
20Static egress classification registry (LLM endpoint patterns)SmallBackend

Deferred to W2

  • AutomationRun persistence and automation_runs collection
  • Topology/run correlation and version tracking
  • Chain fingerprint versioning (ADR-008 Phase 2)
  • Full blast radius traversal beyond first boundary
  • Authority Chains page redesign (not in W1 navigation)
  • "Was → is" risk velocity tables (Sergey's full delta vision beyond refresh delta)

5. Two-Axis Execution Status Model (CISO Recommendation)

The CISO analysis identified a critical refinement: W1 needs two distinct axes, not one:

AxisQuestionValuesSource
execution_proven"Do we have deterministic evidence this executed?"true / falseExistence of linked execution_evidence records
authority_status"Is this authority actively exercised or dormant?"active / dormantRecency of execution within 30-day window

Current doc 06 collapses these into a single binary execution_status: proven | unproven. The W1 UX requires both — the homepage shows "Active Non-Human Execution Identities" (authority_status = active) and "Dormant Authority Identities" (authority_status = dormant), while finding details need to show whether evidence linkage was deterministic (execution_proven).


6. Internal vs External Vocabulary

Internal (code/architecture)External (CISO/W1 UX)
execution_chains collectionAuthority Chains (secondary, not in W1 primary flow)
AutomationTopology (concept)Not exposed to W1 users
AutomationRun (W2 concept)Execution Instance (W2)
ExecutionEvidenceEventExecution Evidence
ExposureUnit (doc 06 concept)Finding row / Exposure
execution_status: proven/unprovenExecution Evidence: Proven / Unproven
risk_cluster (computed)Autonomous Authority Risk Cluster
(unnamed)Standing Authority

The principle: governance language for users, technical language for code.


7. Summary

The 4 analysts converge on these key points:

  1. Doc 06/07's architecture is sound — keep the 4-concept separation internally
  2. W1 demands a different surface — risk clusters, not chains; governance language, not architecture terms
  3. Most of doc 07's naming questions are moot for W1 — the pages they target aren't in W1's flow
  4. The real work is building W1's new views — Homepage with risk clusters, cluster-grouped findings, simplified finding detail
  5. The current platform has all the data building blocks — no schema changes, no new collections; new evaluator rules, new API aggregation, new UI pages
  6. Authority Chains is the consensus internal name (when needed), Executions (30d) is the consensus count label
  7. Risk Cluster concept needs an ADR — it's a new architectural primitive
  8. Standing Authority needs to be named and defined — it's the central W1 governance assertion