Inspector

This page describes the draw.io skill exploration as incomplete, with the diagram never generated and the workflow undocumented. However, the related pages confirm the diagram was successfully created in the same session window (2026-05-21), the draw.io desktop app was installed via Homebrew, and the full Karazhan raw pipeline was diagrammed and rendered. The "Open Threads" and "Next Time, Start Here" sections are now obsolete. This page should be updated to reflect completion, or archived as a superseded draft.

The assets.md entry presents the setup manual as the current reference artifact, but the manual itself has been marked superseded and replaced by a new canonical location. The assets table should be updated to reflect that this asset has been superseded, pointing to the new Context Engine starter kit location instead.

This page records an open thread that has since been resolved. The companion page `karazhan-raw-pipeline-diagram.md` shows the diagram was built in the same session period, mapping the full Karazhan raw-file pipeline. The "Next Time Start Here" action items are now obsolete. This page should be updated to reflect completion or archived as a resolved session log.

The paperclip.md page describes Phase 4 as still in planning and lists open threads for it, but the vault contains a dedicated session log explicitly titled "Phases 1–5 Complete" from the same date (2026-05-21), indicating the migration was finished in full. The page needs its migration status updated from "Phase 4 in planning" to "Phases 1–5 complete," and the Phase 4 open threads section should be revised or archived to reflect that cutover is done.

The QA artifact asserts the security dimension passed with zero open findings, but the security.md artifact for the identical build (776a24d9) and slice (S-prose) explicitly records a finding and a verdict of "findings_open." The QA page's security score and notes directly contradict the evidence in the security review. The QA record should be revised to reflect the actual security verdict, or re-run after the finding is resolved.

The page records session-time stats (103 workouts, 12-week streak) as "current stats," but these are now superseded by the live auto-updating workout-stats.md file. However, since this is a session-notes page capturing a point-in-time snapshot on 2026-05-23, those numbers are historically accurate for that date — they are not false claims. The more meaningful stale signal is that the "Next session — start here" checklist item "Swap / skip per exercise in active logging (not yet built)" has since been built, as the 2026-06-03 context engine audit session confirms "live steering (swap/add/remove)" was designed and built. The page should have its checklist item marked complete or the page archived as a historical session log rather than an active task reference.

The design page lists Iris as a migration target with a specific file path and migration action, but the Phase 0 Verification Findings page explicitly states "No Iris gateway integration needed — Iris does not call Anthropic directly" and "Direct Anthropic SDK calls found: 0." The migration table row for Iris should be removed or annotated as N/A, and the Critical Files section's reference to `hq/iris/iris/core/llm.py` and `hq/iris/iris/config/settings.py` should be updated to reflect that no migration is needed for Iris.

The page describes gsd-pi as having dropped Claude as a runtime dependency, but dash-luxor-agents-m001-complete.md explicitly records GSD v2.74.0 running on the pi SDK with Claude Sonnet 4.6 as the active model for all agent work. This claim is false as of the M001 build (April 2026). The page should be updated to reflect that GSD still depends on Claude, or clarify whether a later version broke that dependency.

This session log was written on 2026-05-29 when only Phases 0–2 were complete and Phases 3–4 were still open. A subsequent page (hq-phase3-phase4-complete-2026-05-31) indicates those phases were completed by 2026-05-31. The "Progress" line and "Open Threads" section now describe work that has already been finished. The page should be updated to reflect that all five phases are complete, or a note should be added pointing to the completion log.

This exploration/session-notes page captured an intermediate state where JARVIS-neon was the leading candidate, but a subsequent session on the same day (2026-05-28) concluded with Minority Report-neon being locked as the final design (saved as `FINAL.html`). The "Open Threads" and "Next Session" sections also reflect this unresolved state, which has since been resolved in favor of MR-neon. The page should be updated or annotated to note that the decision was superseded by the handoff document.

The page places Bridget firmly within Iris's R&D organizational structure as a skill-level component, but the design doc explicitly states she has been promoted out of that role and rebuilt as an independent homogenized agent. The design doc also notes she "was previously slated to retire with the R&D division" — implying the R&D framing on this page is now superseded. This page needs a significant rewrite to reflect Bridget's new architecture, location (`hq/bridget/`), and relationship to Sensei rather than Iris.

The pickleball-log describes the DUPR rating system implementation as a future goal, but the related session notes confirm it was built and a significant calibration bug was already identified and fixed. The "future patch" framing is no longer accurate — the system is live. The Rating & Tracking section should be updated to reflect that a DUPR-style rating system is in place and operational, replacing the "future patch" language.

The design spec was written with port 8790, but when Sensei was actually built on 2026-05-23, it was deployed at port 8794. Every reference to port 8790 in the spec (the header, the launchd plist, the API surface documentation, and the environment variables) is now incorrect. The spec should be updated to reflect port 8794, or a note should be appended recording that the port was changed during implementation.

This implementation plan describes future work across Phases 1–5 with unchecked checkboxes, gate reviews pending, and warnings like "Do not start Phase 2 until Phase 1's findings doc shows a GO decision." However, the session log in personal/hq-paperclip-migration-phases-1-5.md confirms that all five phases were completed in a single ~5.5-hour session on 2026-05-21, including Phase 4 cutover, Phase 5 governance activation, and final acceptance. The plan should be archived or marked as completed to reflect that it is a historical record, not an active execution guide.

This page was captured as a forward-looking idea on 2026-06-09, but the build log from the same date confirms the MCP server was already built and the work done was just gap-filling (adding `vault_list_recent` and a connector). The "What Needs Building" section is now largely obsolete. The page should be updated to reflect that the MCP server exists, note what was actually built, and either archive the page or convert it to a completed-idea record.

The security page shows "findings_open" as its verdict with 1 INFO-level finding, but the QA page (produced 6 seconds later) records "0 open findings" for the Security dimension and awards a perfect score of 100. This contradiction suggests the security verdict was superseded — either the finding was resolved or the INFO classification was confirmed non-blocking and the security page was never updated to reflect closure. The page should be updated to reflect a resolved or closed verdict consistent with the QA record.

The Problem Context section still describes iMessage as a standalone Air script, and Phase 2 still frames the Mac mini migration as future work. The `personal/messages-sync.md` page (updated the same day this doc was created) confirms the migration is already done. The Phase 2 block should be updated to reflect it has shipped, and the Problem Context description of iMessage should be corrected to reflect the current Karazhan adapter setup.

This page documents restoring ccusage as the status line tool on 2026-06-04, but three days later a subsequent session explicitly reversed that decision — ccusage was rejected and replaced with a native implementation. The "If This Breaks Again" troubleshooting steps and the decision to use `npx -y ccusage statusline` are now actively misleading. The page should be archived or annotated with a superseded notice pointing to the 2026-06-07 session log.

The page body still reads as though Dash_Luxor_Agents is the live, authoritative project description, but the page's own Related Pages section explicitly declares it superseded and renamed. The page needs to be either archived with a clear deprecation header pointing to the Routines Manager page, or converted to a historical record with its status updated to reflect that it was the pre-rename project brief.

The smaug.md page claims Smaug is live and in soak testing, but the Iris HQ Audit Final Report (conducted 2026-05-09, ingested 2026-05-14) identified Erebor as non-functional — a P1 finding. The status field and any "Live" language should be updated to reflect the audit's finding, and the page likely needs a note referencing the audit and its implications for Erebor/Smaug's operational state.