c124 review list: the cycle stories, for later

2026-10-26. Everything the story drafts (design c1–c124, and 89 sister-project cycles in section 8) and the completion audit left for Mark to confirm, change or rule on. The stories are live as auto drafts. Nothing here blocks them; each item makes one story more accurate.

Reply in shorthand, e.g. "ok c78 c79; drop c65 Peterson; c14 remove; c120 use c7 before".

1. Closing cycles

The audit table at the bottom of the survey page has 37 Active entries that look ready to close; completion-audit.md has the reasons. On top of the audit:

• c121: marked complete, medium confidence. Hold it open until c123 kicks off?
• c8 Sharing Design Philosophy: park with a PICKUP.md, or close as superseded? Planned in February, never recorded, still framed around data governance.
• c14: the folder holds only the Roesler, Woods & Feil PDF, already in brand/foundational/sources. Remove it, or record it as superseded by c117?
• c73: a two-word note. Close as abandoned?
• c74: a good question about how we build parts, never answered. Its model codebase (app.qpoint.io) is now stale. Park or close?
• c75: CyclesCard became a design-site component that three pages use, but it never moved to the layer. Park, or close as done locally?
• c83, c86: orphans in neither index. c86's prompt file is empty. Close both?
• c97 duplicate number: c97.harness-chip has a prompt and a plan, and was never built. Renumber or merge?
• c11: "paused" in HISTORY. Is it really superseded by the www Q2 redesign? Nothing links the two.
• c13: paused, and its Outcome is blank.

2. Outside credits to confirm (no source, no credit)

cited means the record names the source. matched means the draft linked it by date or topic. topic only means the credit quotes the call but ties to no decision.

Not credited, on purpose:

• AMZ on c62: the Sep 2 session planned design-partner pilots, but no single observation says so.
• Sarah on c51: the "real user" test on her Mac has no record that it ran.
• c77's outside quotes: they come from the archived phase-1 market-analyst corpus.
• The unnamed designer on c101, and the outside colleague on c112.

3. People to confirm

• Cam Esdaile: Cisco, build partner, since 2026-02-24, the first build partner on record. Is the relationship right?
• Jon Friesen: team, engineering. The records only say "Jon" (qcontrol's owner, who hosted qmap); the full name was matched from qcontrol commit authors.
• Kamal Nasser: team. Credited on c80 for the fix in qcontrol PR #714, which the c80 record doesn't name.
• c1–c6 credits come from git authorship only, because those records name nobody.

4. Images to check

• Early heroes are the 2026-08-15 backfill. c1–c75 mostly show pages as they stood in August, filed by filename prefix, not at the time of the work. Each caption says so.
• Design → build pairs: c15 and c19–c22 pair bob-wire's March policy designs (cropped) with the built parts, and c54 pairs a Figma frame with its page. Only c19's record cites the wireframe.
• Borrowed frames:
  • c94, c95 and c98 use c92 captures from just before and after each change.
  • c118's "after" is c121's first-morning shots.
  • c82 uses c81's Fleet Overview.
  • c19's after is c15's capture (taken after c22's change).
• Sister-project heroes: c43 (qmap's map) and c45 (bob-wire's policy page). Set them to null if heroes should be design captures only.
• c99: the hero is from the archived Opus build. The Fable app that shipped has no capture in the cycle folder.
• c120: c7-method.png still shows the pre-pivot /method text, a true "before". Add it, with a fresh capture of /method as the after?
• c35: the before → after isn't the same subject (Jon's terminal against the landing preview). Nothing records whether the welcome screen shipped.
• Small faults:
  • c31's mock shows "NaN" under Scans Today.
  • One of c68's tagged captures is from an earlier concept round.
• Manifest fixes:
  • c57-agent-pawn.png has no cycle field.
  • c54-policy-enforcement.png and c56-compliance.png are tagged to cycles that didn't make them.
  • c93-column-picker-full.png isn't in the manifest.
• Captures worth taking: they would turn words-only stories into pictures.
  • c105: the cost-center mock.
  • c113: demo.qpoint.io.
  • c111: /cycles/weeks.
  • c114: the composed pages after Oct 9.
  • c117: /sources.
  • c122: /library.

5. Records that disagree

• c10: HISTORY.md and CYCLE.md describe different work.
• c22: CYCLE.md says the Remediate button and the samples footer render; HISTORY.md says they were commented out. The story follows HISTORY.
• c50: HISTORY says completed the day it was created; the record still says in progress, with a "Still open" list.
• Stale status lines:
  • c40 and c48 say Active and c43 says Planned, though all three are closed.
  • c79's CYCLES.md note says the layer release is pending (0.9.11 shipped).
  • c81's says "No pages built yet".
  • c62's says "visibility stream active".
  • c92's is a paragraph-long changelog.
• c47: the record calls its brief brief.md; the file is pixel-pawns-brief.md.
• c49: the record says the deploy ships committed work only; since 2026-10-18 it builds the working tree.
• c37: the record says network sessions show live, but its own raw capture lists no network events.
• c65 and c70: listed Active, but their records say Completed.
• Stale PICKUP notes: c114's says parked on 2026-10-09; c121's is the finished Wave 3 hand-off.
• c76 and c82: listed Active with no CYCLE.md.

6. Drafted structure, not from the record

• Decision ids: many records have no numbered Decisions, so the story ids are placeholders in the order the rulings appear: c51, c54, c58, c59, c62, c63–c72, c77, c79, c81, c82, c84, c85, c87, c88–c93 and c118. Where a record didn't say who decided, the decision is attributed to Mark. Rename or drop as needed.
• Modes: most early records have no Mode line, so modes are marked "(inferred)" or "none recorded".
• Approximate dates:
  • c1–c12 (records say "February/March");
  • c17, c19, c21, c22 and c25 (from first commits);
  • c45 (first commit 2026-05-20; the bob-wire page is dated 2026-05-04);
  • c73 (git says 09-01; the folder says 08-15);
  • c76 (the prompt file's date).

7. market-analyst data fixes

• Verbatim quotes that don't match their observations: 5b4bae13#0 and 1668d68e#3. Both are excluded from the stories.
• Call 09ecfede is dated 2026-09-03 but linked to c90 mock pages that were built from 09-16. Is the date wrong?
• Cam's nine phase-1 sessions (in briefs/cam-dashboard-directives.md) aren't in all-calls.json, so c55/c62 cite the brief instead (brief-cam-dashboard-directives#…).

8. Sister projects (qmap, qflow, bob-wire, market-analyst, qpixel, www, app)

89 stories, drafted 2026-10-26. 48 of the cycles are open, and each open one carries a pickup: the next step, what it waits on, and the last activity. Most of these records name nobody, so Mark's credit rests on git authorship, and each credit's detail says so. Modes are inferred throughout.

Outside credits to confirm

Not credited:

• "devindudeman" made copy edits to www's home and QControl pages on 2026-05-18. It isn't confirmed that this is Devin Bernosky.

Ready to close (paperwork only)

Each project's own CYCLES.md and HISTORY.md, not design's:

• qflow: c8, c10.
• qmap:
  • c18–c20 (in neither index);
  • c17 (superseded by c19);
  • c21–c24, which need Outcomes;
  • c22 says Completed but was never moved to HISTORY.
• bob-wire:
  • c1, c3, c4, c9, c10;
  • c11–c17, c19, c21;
  • c2 (superseded by c3).
• market-analyst: c4, c7, c8, c11.
• www: all six. The work merged in PRs #58 and #61 in May, and the records were never updated. c5's banner was removed 2026-07-30.
• qpixel: c3 (an empty folder, replaced by c4); c4 and c5 once they're committed.
• app: c1, c2. app.qpoint.io is no longer used.

Work that exists only on disk (never committed)

• qflow: c12's experiment.
• qmap: c26, c27, c28, c23's model loader and some c25 edits.
• bob-wire: c17's pages and data, c20's pages, and the whole c21 folder.
• qpixel: c4 and c5, inside the uncommitted library rewrite.

Records that disagree

• Dates a year off: qmap c1–c16 records say 2025, but git puts the work in early 2026 (the repo began 2026-01-29). The stories keep the record dates and flag them.
• Copied records: qmap c7's record copies c6's word for word, and c13's copies c12's.
• market-analyst c7: its PICKUP note says the schema and prompt work wasn't started, but a 2026-03-17 commit did it.
• market-analyst c12: the record is a step behind today's commit (33 candidates, not 30).
• bob-wire c2: the number is used twice. The cycle folder holds the dataflows prompt; the c2 pages are an unrelated landing mockup.
• bob-wire c14: the planned lens names don't match what was built.

Images to fix or take

• Bad captures (captions corrected in the drafts; the manifests are unchanged):
  • qflow c0's nine frames are all black;
  • qmap c17 shows empty scenes;
  • six of the nine qmap c20 frames show only the end points;
  • three qmap c24 scenarios look identical;
  • qmap c27's usage-sized frame matches the flat roster;
  • qmap c23's "interns" frame shows Helpers;
  • qmap c28's rug reads "Session #undefined".
• market-analyst c8: its four embedded screenshots were never saved.
• Worth capturing:
  • market-analyst c12's /problems pages, as an after for the c2 Insights grab;
  • bob-wire c21's three design-site mocks (its hero borrows a c92 capture);
  • bob-wire c1, c6 and c9's mockups;
  • www c2's built page (the hero is the Illustrator design).
• Left out on purpose:
  • the PocketOS case-study PDF (marked confidential);
  • bob-wire c18 and c19's precedent screenshots (other vendors' products);
  • bob-wire c4's prototype shots that show an email address and account ids.