{"reviews":{"source":{"verdict":"pass","summary":"All 33 claims are supported within their declared scope. Live primary-source passages and the saved ResearchNode original were inspected. The seven original modules, embedded exercises, checklist and quiz objectives have documented adaptations without material omissions. Diagram geometry, labels, editable XML and explanations agree. Calculations are correct. Only the coordinator’s eleven Python fixture decisions are treated as executed; other examples remain illustrative. No private material or material unsupported generalization was found. This is source-support review, not independent factual verification.","issues":[],"claims":[{"id":"C1","verdict":"supported","reason":"The actual notebook and coverage mapping retain all seven modules, with explicit explanations for replacing operational examples and interactive controls."},{"id":"C2","verdict":"supported","reason":"Host/client/server responsibilities and self-contained requests match the live [architecture specification](https://modelcontextprotocol.io/specification/2026-07-28/architecture)."},{"id":"C3","verdict":"supported","reason":"Both historical lifecycle pages specify initialization, the initialized notification, capability-aware operation and transport termination: [2024](https://modelcontextprotocol.io/specification/2024-11-05/basic/lifecycle), [2025](https://modelcontextprotocol.io/specification/2025-11-25/basic/lifecycle)."},{"id":"C4","verdict":"supported","reason":"The versioning page, reopened through its lifecycle redirect, confirms per-request versioning, discovery, supported-version errors and optional dual-era implementations. [Versioning](https://modelcontextprotocol.io/specification/2026-07-28/basic/versioning)."},{"id":"C5","verdict":"supported","reason":"Framing, diagnostic separation and stdin-closure shutdown match the live [stdio specification](https://modelcontextprotocol.io/specification/2026-07-28/basic/transports/stdio)."},{"id":"C6","verdict":"supported","reason":"POST delivery, request-scoped SSE, removal of sessions/GET streams, subscriptions, cancellation and endpoint controls match the scoped [HTTP revision](https://modelcontextprotocol.io/specification/2026-07-28/basic/transports/streamable-http)."},{"id":"C7","verdict":"supported","reason":"All six directed edges agree across screenshot, exported path geometry, embedded cells, editable XML, alt text and prose. The decomposition is explicitly application-specific."},{"id":"C8","verdict":"supported","reason":"MCPServer, tool decorators, default stdio and run-time transport options match the live [Python SDK documentation](https://py.sdk.modelcontextprotocol.io/run/)."},{"id":"C9","verdict":"supported","reason":"Factory serving, registration and external security middleware match the reopened TypeScript [stdio](https://ts.sdk.modelcontextprotocol.io/v2/serving/stdio), [HTTP](https://ts.sdk.modelcontextprotocol.io/v2/serving/http.html) and [tools](https://ts.sdk.modelcontextprotocol.io/v2/servers/tools.html) documentation."},{"id":"C10","verdict":"supported","reason":"Static inspection confirms matching decision order and all eleven fixture outcomes for the scoped inputs. Trusted-input assumptions and absent operational enforcement are clearly disclosed."},{"id":"C11","verdict":"supported","reason":"Invisible-character risks, legitimate-text preservation and the limits of delimiters match the inspected [OWASP guidance](https://cheatsheetseries.owasp.org/cheatsheets/LLM_Prompt_Injection_Prevention_Cheat_Sheet.html)."},{"id":"C12","verdict":"supported","reason":"The live [Tools specification](https://modelcontextprotocol.io/specification/2026-07-28/server/tools) describes subscribed list-change delivery. Reapproval is correctly identified as application policy."},{"id":"C13","verdict":"supported","reason":"The resource/aud distinction and separate downstream-token requirement match [RFC 8707](https://www.rfc-editor.org/rfc/rfc8707.html) and the cited [authorization revision](https://modelcontextprotocol.io/specification/2025-11-25/basic/authorization)."},{"id":"C14","verdict":"supported","reason":"Per-client consent matches [proxy guidance](https://modelcontextprotocol.io/docs/2025-11-25/tutorials/security/security_best_practices). The saved document and revision match the package; its impersonation discussion agrees with the [live original](https://blog.modelcontextprotocol.io/posts/client_registration/)."},{"id":"C15","verdict":"supported","reason":"Avoiding OS commands and combining structured arguments with semantic validation match [OWASP command-injection guidance](https://cheatsheetseries.owasp.org/cheatsheets/OS_Command_Injection_Defense_Cheat_Sheet.html)."},{"id":"C16","verdict":"supported","reason":"Both original command arrays put the branch variable after --. The [Git manual](https://git-scm.com/docs/git-log) identifies that position as path operands and confirms the stated manual version."},{"id":"C17","verdict":"supported","reason":"Identifier mapping, component-aware containment and pre-open symlink races are explicitly covered by the live [path guidance](https://cheatsheetseries.owasp.org/cheatsheets/Path_Traversal_Prevention_Cheat_Sheet.html)."},{"id":"C18","verdict":"supported","reason":"The original returns a URL after a separate DNS lookup, then invokes httpx independently. The bounded inference and redirect limitations align with [OWASP SSRF guidance](https://cheatsheetseries.owasp.org/cheatsheets/Server_Side_Request_Forgery_Prevention_Cheat_Sheet.html)."},{"id":"C19","verdict":"supported","reason":"All four original vendor-risk objectives survive as review tasks and evidence requests, without unsupported vendor-specific incident or mitigation-success claims."},{"id":"C20","verdict":"supported","reason":"The seven layer names and cross-cutting L6 match the inspected [CSA MAESTRO article](https://cloudsecurityalliance.org/blog/2025/02/06/agentic-ai-threat-modeling-framework-maestro). Original substituted names are accurately identified."},{"id":"C21","verdict":"supported","reason":"The six categories and their meanings match the readable [Microsoft STRIDE table](https://learn.microsoft.com/en-us/azure/security/develop/threat-modeling-tool-threats). MCP examples are clearly adaptations."},{"id":"C22","verdict":"supported","reason":"Attribution, date and three capabilities match [Willison’s original article](https://simonwillison.net/2025/Jun/16/the-lethal-trifecta/). Numerical risk calibration is not attributed to it."},{"id":"C23","verdict":"supported","reason":"All products are correct: 16, 15, 20, 12 and hypothetical 10. Scores, mappings and control updates are explicitly synthetic; register fields and exercises are present."},{"id":"C24","verdict":"supported","reason":"Each denial and check-omission counterfactual follows from the supplied policy. The five original scenarios remain recognizable, with explicit limits on what the substitutions demonstrate."},{"id":"C25","verdict":"supported","reason":"Both illustrative suites contain the stated positive, negative, malformed-input, non-mutation and repeatability checks. The separate coordinator execution supports only the eleven standalone Python fixture decisions."},{"id":"C26","verdict":"supported","reason":"The parametrization API and example structure match [pytest documentation](https://docs.pytest.org/en/stable/how-to/parametrize.html); no framework execution is claimed."},{"id":"C27","verdict":"supported","reason":"Table-driven testing is supported by the [Vitest API](https://vitest.dev/api/test). Mixed malformed-value cases are consistent with the inspected [maintainer implementation](https://raw.githubusercontent.com/vitest-dev/vitest/v3.2.4/packages/runner/src/suite.ts); runtime success remains unclaimed."},{"id":"C28","verdict":"supported","reason":"All fifteen original checklist objectives are represented under the six actual script domains. The notes template preserves study-recording objectives without claiming browser persistence."},{"id":"C29","verdict":"supported","reason":"The complete embedded original confirms five stored scenario narratives, six quiz items and fifteen checks across six domains. Adaptation disclosures accurately distinguish original inspection, synthetic execution and unexecuted examples."},{"id":"C30","verdict":"supported","reason":"Topology matches the architecture source; envelopes, IDs, exclusive result/error members and notification behavior match the [JSON-RPC specification](https://www.jsonrpc.org/specification)."},{"id":"C31","verdict":"supported","reason":"Legacy subscription methods and modern resourceSubscriptions filtering match the [migration guide](https://py.sdk.modelcontextprotocol.io/migration/) and [Resources specification](https://modelcontextprotocol.io/specification/2026-07-28/server/resources). Delivery checks are presented as a plan."},{"id":"C32","verdict":"supported","reason":"Binding values separately from query structure matches [OWASP SQL guidance](https://cheatsheetseries.owasp.org/cheatsheets/SQL_Injection_Prevention_Cheat_Sheet.html). Object authorization is appropriately treated as a separate design check."},{"id":"C33","verdict":"supported","reason":"Discovery, dispatch and distinct protocol/tool errors align with the [Tools specification](https://modelcontextprotocol.io/specification/2026-07-28/server/tools); ID correlation matches JSON-RPC. The exercise claims no executed transport coverage."}],"package_sha256":"b6a360757ef39d364327d0d5d1d86dd1baa0ea05c5e2d67569b5bf1631179d11","reviewed_at":"2026-11-19T22:10:58.978171+01:00"},"visual":{"verdict":"pass","summary":"Reviewed the supplied manifest-bound desktop, 390px and 320px captures, all panel-scroll positions, question states and 20 print pages against the approved references and supplied content. No material visual or semantic inconsistencies found. Dark styling, syntax colours, rounded question badges, full numbered URLs and readable spacing are retained. Diagram arrows and TB labels match the text alternative. Dense content scrolls within panels; recorded checks report no page overflow and successful keyboard, copy, navigation and download behavior. Print includes all disclosure answers and complete content. Browser results are supplied evidence, not independently rerun; factual accuracy is not certified.","issues":[],"claims":[],"package_sha256":"b6a360757ef39d364327d0d5d1d86dd1baa0ea05c5e2d67569b5bf1631179d11","reviewed_at":"2026-11-19T22:10:58.978214+01:00"}},"checks":{"package_sha256":"b6a360757ef39d364327d0d5d1d86dd1baa0ea05c5e2d67569b5bf1631179d11","passed":true,"widths":[1440,390,320],"code_checked":true,"code":[{"language":"python","syntax":"passed","sha256":"1300641be32dbc5e57323e575cbf32f9927d7a3f61eefe82e19f01e1271ea76b"},{"language":"typescript","syntax":"illustrative","sha256":"ea79da6cbc3bb61d5064978783b54e4e5de81a1641a24bac72684f0b33666e0a"},{"language":"python","syntax":"passed","sha256":"3bb015d5b2b551f590bd29464e03b2618cb3a89311a81cbdbf1dce7ce3f833cc"},{"language":"typescript","syntax":"illustrative","sha256":"735bd812684fb71a69f0f4eabf4daffa394a0cc5df2518b7311e47f249a93fd6"}],"execution":[{"file":"lab.py","sha256":"1894e9e3aad2338d41f80383220e29cff74fe8b29593ec4dfb647113008a11ec","returncode":0,"output":"11 fixture decisions matched expected results\n","environment":"CPython 3.12 isolated container; no network, no host mounts; 128 MiB; 1 CPU"}],"drawio":["boundaries.drawio"],"print":"rendered"},"published_at":"2026-11-19T19:58:29.293Z"}