kf-2026.10.18-6

Graph report

Revision 47a3ad8; 199 nodes, 502 edges, 8 communities.

Nodes by type

• Concept: 28
• Essay: 38
• Initiative: 10
• Organization: 6
• Playbook: 3
• Practice: 16
• Source: 51
• Standard: 8
• Theme: 5
• Tool: 34

Most connected notes

• Adopting MCP in a bank: the four risks to solve first (30)
• Architecture that runs: why AI-era governance must be code (30)
• The AI-native SDLC: when typing code stops being the bottleneck (29)
• Model is a tool, harness is the strategy (26)
• Agent-first data architecture: five principles (24)
• AI-Augmented Software Development Lifecycle (23)
• Spec-driven development for regulated industries: audit the contract, not every line (19)
• Self-describing knowledge: why the Open Knowledge Format matters, and what enterprises must add (19)
• Inside an Architecture-as-Code platform: a C4 walkthrough (19)
• Capture the why: the knowledge layer enterprise agents are missing (18)

Orphans

• none

Semantic layer

• Extractor: semantica:llm/pattern
• Entities: 92
• Extractor note: llm pass: 51 notes in 151s, slowest governed-ai/adopting-mcp-in-a-bank-the-four-risks-to-solve-first.md at 7s
• Extractor note: semantica warned: fastembed not available. Install with: pip install 'semantica[embeddings-local]'. Using fallback embedding method.
• Extractor note: resolution: 243 names -> 234 entities; spec = specs (spelling); agents = agent (spelling); pipeline = pipelines (spelling); AI agent = AI agents (spelling); OKF = Open Knowledge Format (acronym); MCP = Model Context Protocol (acronym); AI-generated code = AI-written code (model); Living documentation = living document (model); maker and checker = maker/checker controls (model)
• Extractor note: kept 92 of 234 resolved entities (two or more notes, or a named standard, tool or organisation)
• Extractor note: relation confidences returned: 0.75, 0.75, 0.75, 0.75, 0.75, 0.75, 0.75, 0.75, 0.75, 0.75, 0.75, 0.75, 0.75, 0.75, 0.75, 0.75, 0.75, 0.75, 0.75, 0.75, 0.75, 0.75, 0.75, 0.75, 0.75, 0.75, 0.75, 0.75, 0.75, 0.75
• Extractor note: typed relations: 45 kept, none dropped
• Extractor note: relation pass: 182s
• Typed relations proposed: 38 new (of 45 found; the rest repeat an existing edge)
  • ai-native-sdlc/ai-augmented-software-development-lifecycle.md applies ai-native-sdlc/governing-coding-agents-in-a-bank-a-12-point-checklist.md (0.75): "Coding agents brought into a large bank's software lifecycle behind one governed platform: curated context, maker/checker controls and a single gateway for ever"
  • ai-native-sdlc/ai-augmented-software-development-lifecycle.md applies ai-native-sdlc/three-architectures-for-ai-coding-agents-and-how-enterprises.md (0.75): "Coding agents brought into a large bank's software lifecycle behind one governed platform: curated context, maker/checker controls and a single gateway for ever"
  • ai-native-sdlc/governing-coding-agents-in-a-bank-a-12-point-checklist.md applies ai-native-sdlc/beyond-vibe-coding-specs-as-executable-artifacts.md (0.75): "10. Use specs for high-risk changes"
  • ai-native-sdlc/governing-coding-agents-in-a-bank-a-12-point-checklist.md applies governed-ai/architecture-that-runs-why-ai-era-governance-must-be-code.md (0.75): "6. Put policy in the pipeline"
  • ai-native-sdlc/spec-driven-development-for-regulated-industries-audit-the.md builds_on ai-native-sdlc/three-architectures-for-ai-coding-agents-and-how-enterprises.md (0.75): "Making the specification the contract an AI must fulfil can, provided the spec stays anchored to the code and machines check conformance."
  • ai-native-sdlc/spec-driven-development-for-regulated-industries-audit-the.md applies ai-native-sdlc/ai-augmented-software-development-lifecycle.md (0.75): "For complex work in regulated sectors, prompt-to-code cannot pass an audit."
  • ai-native-sdlc/the-ai-native-sdlc-when-typing-code-stops-being-the.md builds_on ai-native-sdlc/governing-coding-agents-in-a-bank-a-12-point-checklist.md (0.75): "The real investment is trusted knowledge and controls that live inside the engineering flow, not an approval step bolted on at the end."
  • ai-native-sdlc/the-ai-native-sdlc-when-typing-code-stops-being-the.md builds_on context-engineering/context-fragmentation-why-your-agents-keep-starting-from.md (0.75): "Verification, trusted context and human judgment are now the scarce resources, and teams that redesign their lifecycle around them will pull away from teams tha"
  • context-engineering/living-documentation-and-the-human-trust-gap.md applies data-ai-platforms/capture-the-why-the-knowledge-layer-enterprise-agents-are.md (0.75): "The harder, more valuable problem is capturing why it exists."
  • context-engineering/passive-capture-knowledge-that-writes-itself-with-humans-who.md builds_on ai-native-sdlc/the-ai-native-sdlc-when-typing-code-stops-being-the.md (0.75): "Capture it passively, distill it, and keep a human approval gate before anything becomes trusted."
  • context-engineering/passive-capture-knowledge-that-writes-itself-with-humans-who.md applies ai-native-sdlc/ai-augmented-software-development-lifecycle.md (0.75): "Capture it passively, distill it, and keep a human approval gate before anything becomes trusted."
  • context-engineering/self-describing-knowledge-why-the-open-knowledge-format.md builds_on context-engineering/not-another-data-lake-context-needs-curation-not-collection.md (0.75): "Regulated enterprises still need to add ownership, classification and a policy for how much trust an agent needs before it acts."
  • context-engineering/self-describing-knowledge-why-the-open-knowledge-format.md builds_on ai-native-sdlc/the-ai-native-sdlc-when-typing-code-stops-being-the.md (0.75): "Knowledge for the AI era should be plain, portable and self-describing: Markdown people can read, structure agents can parse, and metadata that says who owns it"
  • data-ai-platforms/capture-the-why-the-knowledge-layer-enterprise-agents-are.md refines context-engineering/self-describing-knowledge-why-the-open-knowledge-format.md (0.75): "Durable, governed context about the why is the real foundation for useful agents."
  • data-ai-platforms/enterprise-machine-learning-platform.md applies data-ai-platforms/from-mlops-to-ai-engineering-what-industrializing-ml-still.md (0.75): "An enterprise MLOps platform that moved 360 data scientists onto one governed environment, saved $25M in license costs over 3 years and cut model deployment fro"
  • data-ai-platforms/from-mlops-to-ai-engineering-what-industrializing-ml-still.md builds_on governed-ai/enterprise-generative-ai-architecture.md (0.75): "The lessons of MLOps still hold for generative and agentic AI: industrialize the path to production, measure what happens after launch, and give risk teams the "
  • data-ai-platforms/governance-belongs-in-the-catalog.md builds_on data-ai-platforms/agent-first-data-architecture-five-principles.md (0.75): "When business meaning, data products and relationships live in one open catalog, agents stop guessing at tables and governance travels with the data."
  • governed-ai/a2a-and-mcp-complementary-not-competing.md builds_on governed-ai/adopting-mcp-in-a-bank-the-four-risks-to-solve-first.md (0.75): "MCP connects agents to tools; Agent2Agent connects agents to each other."
  • governed-ai/adopting-mcp-in-a-bank-the-four-risks-to-solve-first.md builds_on ai-native-sdlc/governing-coding-agents-in-a-bank-a-12-point-checklist.md (0.75): "Four risk areas a regulated institution should resolve before agents get tools."
  • governed-ai/adopting-mcp-in-a-bank-the-four-risks-to-solve-first.md builds_on data-ai-platforms/capture-the-why-the-knowledge-layer-enterprise-agents-are.md (0.75): "Auditability: reconstruct the decision behind the call"
  • governed-ai/adopting-mcp-in-a-bank-the-four-risks-to-solve-first.md applies ai-native-sdlc/ai-augmented-software-development-lifecycle.md (0.75): "Trust is still an architecture problem: identity, policy, audit and a governed server catalog decide whether a bank can use it safely."
  • governed-ai/architecture-as-code-control-plane.md builds_on ai-native-sdlc/the-ai-native-sdlc-when-typing-code-stops-being-the.md (0.75): "An executable, policy-aware governance engine that checks designs automatically before they reach production, so architecture keeps pace with AI-speed delivery."
  • governed-ai/architecture-as-code-control-plane.md applies ai-native-sdlc/ai-augmented-software-development-lifecycle.md (0.75): "An executable, policy-aware governance engine that checks designs automatically before they reach production, so architecture keeps pace with AI-speed delivery."
  • governed-ai/architecture-as-code-reference-platform-open-prototype.md applies governed-ai/model-is-a-tool-harness-is-the-strategy.md (0.75): "aac.muthub.org: a public, working prototype of an architecture control plane, with catalogs, OPA policies, waivers, an executive scorecard, a CLI and an MCP ser"
  • governed-ai/architecture-as-code-reference-platform-open-prototype.md applies ai-native-sdlc/spec-driven-development-for-regulated-industries-audit-the.md (0.75): "It is modeled in its own notation and passes its own rules."
  • governed-ai/architecture-that-runs-why-ai-era-governance-must-be-code.md applies ai-native-sdlc/ai-augmented-software-development-lifecycle.md (0.75): "Governance has to become something that executes in the pipeline and in the agent's context."
  • governed-ai/designing-an-enterprise-mcp-registry-the-control-point.md builds_on governed-ai/a2a-and-mcp-complementary-not-competing.md (0.75): "A governed registry decides which connections are allowed."
  • governed-ai/designing-an-enterprise-mcp-registry-the-control-point.md applies governed-ai/model-is-a-tool-harness-is-the-strategy.md (0.75): "Build the second before you scale the first."
  • governed-ai/if-you-can-t-prove-why-an-ai-wrote-it-you-can-t-ship-it.md builds_on governed-ai/architecture-that-runs-wh...

kf-2026.10.18-5

Graph report

Revision 3453bac; 200 nodes, 484 edges, 8 communities.

Nodes by type

• Concept: 24
• Essay: 38
• Initiative: 10
• Organization: 7
• Playbook: 3
• Practice: 20
• Source: 51
• Standard: 7
• Theme: 5
• Tool: 35

Most connected notes

• Model is a tool, harness is the strategy (32)
• The AI-native SDLC: when typing code stops being the bottleneck (30)
• Agent-first data architecture: five principles (23)
• Inside an Architecture-as-Code platform: a C4 walkthrough (22)
• Architecture-as-Code Reference Platform (open prototype) (21)
• Architecture that runs: why AI-era governance must be code (21)
• The cost of unmanaged context (19)
• Passive capture: knowledge that writes itself, with humans who approve it (18)
• Self-describing knowledge: why the Open Knowledge Format matters, and what enterprises must add (17)
• Adopting MCP in a bank: the four risks to solve first (17)

Orphans

• none

Semantic layer

• Extractor: semantica:llm/pattern
• Entities: 93
• Extractor note: llm pass: 51 notes in 147s, slowest governed-ai/model-is-a-tool-harness-is-the-strategy.md at 7s
• Extractor note: semantica warned: fastembed not available. Install with: pip install 'semantica[embeddings-local]'. Using fallback embedding method.
• Extractor note: resolution: 236 names -> 226 entities; agents = agent (spelling); pipeline = pipelines (spelling); AI agent = AI agents (spelling); catalog = catalogs (spelling); OKF = Open Knowledge Format (acronym); MCP = Model Context Protocol (acronym); AI-generated code = AI-written code (model); Software Development Lifecycle = SDLC (model); maker and checker = maker/checker controls (model); specification = specs (model)
• Extractor note: kept 93 of 226 resolved entities (two or more notes, or a named standard, tool or organisation)
• Extractor note: relation confidences returned: 0.75, 0.75, 0.75, 0.75, 0.75, 0.75, 0.75, 0.75, 0.75, 0.75, 0.75, 0.75, 0.75, 0.75, 0.75, 0.75, 0.75, 0.75, 0.75, 0.75, 0.75, 0.75, 0.75, 0.75, 0.75, 0.75, 0.75, 0.75, 0.75, 0.75
• Extractor note: typed relations: 51 kept, none dropped
• Extractor note: relation pass: 170s
• Typed relations proposed: 41 new (of 51 found; the rest repeat an existing edge)
  • ai-native-sdlc/ai-augmented-software-development-lifecycle.md applies governed-ai/architecture-that-runs-why-ai-era-governance-must-be-code.md (0.75): "Coding agents brought into a large bank's software lifecycle behind one governed platform: curated context, maker/checker controls and a single gateway for ever"
  • ai-native-sdlc/ai-augmented-software-development-lifecycle.md applies ai-native-sdlc/governing-coding-agents-in-a-bank-a-12-point-checklist.md (0.75): "Coding agents brought into a large bank's software lifecycle behind one governed platform: curated context, maker/checker controls and a single gateway for ever"
  • ai-native-sdlc/governing-coding-agents-in-a-bank-a-12-point-checklist.md applies governed-ai/model-is-a-tool-harness-is-the-strategy.md (0.75): "Govern the tools as well as the models"
  • ai-native-sdlc/spec-driven-development-for-regulated-industries-audit-the.md applies ai-native-sdlc/governing-coding-agents-in-a-bank-a-12-point-checklist.md (0.75): "For complex work in regulated sectors, prompt-to-code cannot pass an audit."
  • ai-native-sdlc/spec-driven-development-for-regulated-industries-audit-the.md builds_on governed-ai/model-is-a-tool-harness-is-the-strategy.md (0.75): "Make the spec the contract, keep it anchored to the code, and let machines prove conformance."
  • ai-native-sdlc/spec-driven-development-for-regulated-industries-audit-the.md applies governed-ai/architecture-as-code-reference-platform-open-prototype.md (0.75): "Make the spec the contract, keep it anchored to the code, and let machines prove conformance."
  • ai-native-sdlc/the-ai-native-sdlc-when-typing-code-stops-being-the.md builds_on governed-ai/architecture-as-code-control-plane.md (0.75): "The real investment is trusted knowledge and controls that live inside the engineering flow, not an approval step bolted on at the end."
  • ai-native-sdlc/the-ai-native-sdlc-when-typing-code-stops-being-the.md applies governed-ai/model-is-a-tool-harness-is-the-strategy.md (0.75): "Code fast, verify hard, and keep the humans in the loop."
  • ai-native-sdlc/the-ai-native-sdlc-when-typing-code-stops-being-the.md builds_on context-engineering/context-fragmentation-why-your-agents-keep-starting-from.md (0.75): "Verification, trusted context and human judgment are now the scarce resources, and teams that redesign their lifecycle around them will pull away from teams tha"
  • ai-native-sdlc/three-architectures-for-ai-coding-agents-and-how-enterprises.md applies ai-native-sdlc/ai-augmented-software-development-lifecycle.md (0.75): "Platform-embedded, sandboxed and evolutionary coding agents each earn trust in a different way."
  • ai-native-sdlc/three-architectures-for-ai-coding-agents-and-how-enterprises.md builds_on governed-ai/if-you-can-t-prove-why-an-ai-wrote-it-you-can-t-ship-it.md (0.75): "Pick agents by how you will verify them, and own the harness that makes any of them safe."
  • context-engineering/not-another-data-lake-context-needs-curation-not-collection.md builds_on context-engineering/the-cost-of-unmanaged-context.md (0.75): "Otherwise you are building the next swamp, with agents swimming in it."
  • context-engineering/passive-capture-knowledge-that-writes-itself-with-humans-who.md builds_on context-engineering/context-fragmentation-why-your-agents-keep-starting-from.md (0.75): "The knowledge we need is already produced as a by-product of work: commits, tickets, incidents, reviews and decisions."
  • context-engineering/passive-capture-knowledge-that-writes-itself-with-humans-who.md applies ai-native-sdlc/the-ai-native-sdlc-when-typing-code-stops-being-the.md (0.75): "Capture it passively, distill it, and keep a human approval gate before anything becomes trusted."
  • context-engineering/self-describing-knowledge-why-the-open-knowledge-format.md refines context-engineering/not-another-data-lake-context-needs-curation-not-collection.md (0.75): "Regulated enterprises still need to add ownership, classification and a policy for how much trust an agent needs before it acts."
  • context-engineering/the-cost-of-unmanaged-context.md builds_on ai-native-sdlc/the-ai-native-sdlc-when-typing-code-stops-being-the.md (0.75): "Context is the asset AI runs on, and most enterprises leave it unmanaged."
  • data-ai-platforms/agent-first-data-architecture-five-principles.md builds_on context-engineering/the-cost-of-unmanaged-context.md (0.75): "Supply context, and less of it"
  • data-ai-platforms/capture-the-why-the-knowledge-layer-enterprise-agents-are.md refines context-engineering/context-fragmentation-why-your-agents-keep-starting-from.md (0.75): "Most enterprise AI pilots hit the same wall: systems record what happened, but the reasoning behind decisions evaporates."
  • data-ai-platforms/capture-the-why-the-knowledge-layer-enterprise-agents-are.md builds_on context-engineering/self-describing-knowledge-why-the-open-knowledge-format.md (0.75): "Durable, governed context about the why is the real foundation for useful agents."
  • data-ai-platforms/cloud-data-analytics-platform.md about data-ai-platforms/controlled-data-movement.md (0.75): "A cloud data platform at USAA processing 25B+ records a day, with an AWS architecture approved for identity, data protection and detective controls."
  • data-ai-platforms/controlled-data-movement.md applies leading-ai-change/data-ai-on-public-cloud.md (0.75): "A configuration-driven, secure data movement pattern built at USAA to move 25B+ records a day, then carried to Wells Fargo as a bank-wide framework for analytic"
  • data-ai-platforms/from-mlops-to-ai-engineering-what-industrializing-ml-still.md applies governed-ai/enterprise-generative-ai-architecture.md (0.75): "The lessons of MLOps still hold for generative and agentic AI: industrialize the path to production, measure what happens after launch, and give risk teams the "
  • data-ai-platforms/governance-belongs-in-the-catalog.md builds_on governed-ai/architecture-that-runs-why-ai-era-governance-must-be-code.md (0.75): "When business meaning, data products and relationships live in one open catalog, agents stop guessing at tables and governance travels with the data."
  • data-ai-platforms/governance-belongs-in-the-catalog.md applies governed-ai/architecture-as-code-reference-platform-open-prototype.md (0.75): "When business meaning, data products and relationships live in one open catalog, agents stop guessing at tables and governance travels with the data."
  • governed-ai/a2a-and-mcp-complementary-not-competing.md builds_on governed-ai/adopting-mcp-in-a-bank-the-four-risks-to-solve-first.md (0.75): "Together they turn single agents into systems, and raise the governance stakes."
  • governed-ai/architecture-as-code-control-plane.md about governed-ai/architecture-as-code-reference-platform-open-prototype.md (0.75): "An executable, policy-aware governance engine that checks designs automatically before they reach production, so architecture keeps pace with AI-speed delivery."
  • governed-ai/architecture-as-code-control-plane.md applies ai-native-sdlc/governing-coding-agents-in-a-bank-a-12-point-checklist.md (0.75): "An executable, policy-aware governance engine that checks designs automatically before they reach production, so architecture keeps pace with AI-speed delivery."
  • governed-ai/architecture-as-code-reference-platform-open-prototype.md applies governed-ai/model-is-a-tool-harness-is-the-strategy.md (0.75): "aac.muthub.org: a public, working prototype of an architecture control plane, with catalogs, OPA policies, waivers, an executive scorecard, a CLI and an MCP ser"
  • governed-ai/architecture-as-code-reference-platform-open-prototype.md applies ai-native-sdlc/ai-augmented-software-development-lifec...

kf-2026.10.18-4

Graph report

Revision 1e59b46; 212 nodes, 358 edges, 14 communities.

Nodes by type

• Concept: 85
• Essay: 38
• Initiative: 10
• Organization: 5
• Playbook: 3
• Source: 51
• Theme: 5
• Tool: 15

Most connected notes

• Agent-first data architecture: five principles (19)
• The AI-native SDLC: when typing code stops being the bottleneck (16)
• Adopting MCP in a bank: the four risks to solve first (13)
• Architecture that runs: why AI-era governance must be code (13)
• Not another data lake: context needs curation, not collection (12)
• Inside an Architecture-as-Code platform: a C4 walkthrough (12)
• Model is a tool, harness is the strategy (12)
• Three architectures for AI coding agents, and how enterprises should choose (11)
• From MLOps to AI engineering: what industrializing ML still teaches us (11)
• MCP (11)

Orphans

• none

Semantic layer

• Extractor: semantica:llm/pattern
• Entities: 112
• Extractor note: llm pass: 51 notes in 157s, slowest governed-ai/designing-an-enterprise-mcp-registry-the-control-point.md at 4s
• Extractor note: semantica warned: fastembed not available. Install with: pip install 'semantica[embeddings-local]'. Using fallback embedding method.

kf-2026.10.18-3

Graph report

Revision 4d5e1c2; 107 nodes, 184 edges, 7 communities.

Nodes by type

• Essay: 38
• Initiative: 10
• Playbook: 3
• Source: 51
• Theme: 5

Most connected notes

• The AI-native SDLC: when typing code stops being the bottleneck (9)
• Architecture that runs: why AI-era governance must be code (9)
• Model is a tool, harness is the strategy (9)
• Agent-first data architecture: five principles (8)
• Adopting MCP in a bank: the four risks to solve first (8)
• Inside an Architecture-as-Code platform: a C4 walkthrough (7)
• Spec-driven development for regulated industries: audit the contract, not every line (5)
• Three architectures for AI coding agents, and how enterprises should choose (5)
• The cost of unmanaged context (5)
• Capture the why: the knowledge layer enterprise agents are missing (5)

Orphans

• none

Semantic layer

• Extractor: none
• Entities: 0

kf-2026.10.18-2

Graph report

Revision 4dacb63; 107 nodes, 184 edges, 7 communities.

Nodes by type

• Essay: 38
• Initiative: 10
• Playbook: 3
• Source: 51
• Theme: 5

Most connected notes

• The AI-native SDLC: when typing code stops being the bottleneck (9)
• Architecture that runs: why AI-era governance must be code (9)
• Model is a tool, harness is the strategy (9)
• Agent-first data architecture: five principles (8)
• Adopting MCP in a bank: the four risks to solve first (8)
• Inside an Architecture-as-Code platform: a C4 walkthrough (7)
• Spec-driven development for regulated industries: audit the contract, not every line (5)
• Three architectures for AI coding agents, and how enterprises should choose (5)
• The cost of unmanaged context (5)
• Capture the why: the knowledge layer enterprise agents are missing (5)

Orphans

• none

Semantic layer

• Extractor: none
• Entities: 0

kf-2026.10.17-1

Graph report

Revision 3b1ddd9; 107 nodes, 184 edges, 7 communities.

Nodes by type

• Essay: 38
• Initiative: 10
• Playbook: 3
• Source: 51
• Theme: 5

Most connected notes

• The AI-native SDLC: when typing code stops being the bottleneck (9)
• Architecture that runs: why AI-era governance must be code (9)
• Model is a tool, harness is the strategy (9)
• Agent-first data architecture: five principles (8)
• Adopting MCP in a bank: the four risks to solve first (8)
• Inside an Architecture-as-Code platform: a C4 walkthrough (7)
• Spec-driven development for regulated industries: audit the contract, not every line (5)
• Three architectures for AI coding agents, and how enterprises should choose (5)
• The cost of unmanaged context (5)
• Capture the why: the knowledge layer enterprise agents are missing (5)

Orphans

• none

Semantic layer

• Extractor: none
• Entities: 0