Concept mockup — structure and visual language only. Content is real but abridged; counts reflect the Jul 12–26 run.
CODERAVEN · DISPATCH · CHRONICLE DIGEST

Dispatch · Jul 12–26

Two weeks of identity work — and six tasks your plans still call open that are already done.

Two weeks almost entirely about identity — and then, underneath it, a quieter question about whether your own records can be trusted.

The work
The Dossier stopped being a document
Your Advisory Contract now injects into expert meetings, Mira, and Foundation generation. 14 of 26 commits touched it.
The catch
Your checkboxes are wrong in both directions
Six of eight tasks marked open are already done — and one plan's store says 0/17 while its markdown says 16/17.
The risk
927 archived records rest on two try? calls
The data is intact — I checked every id. But nothing would tell you if it stopped being.

Sessions per day

38 sessions across 15 days — but only 11 of those days saw any work at all.
By weekday

Weekends carried 26% of the fortnight — the sturdiest signal here. Wednesday's 34% is one day, not a habit: the other Wednesday was empty. Two samples per weekday is too thin to call a rhythm; it gets real once a few dispatches stack up.

View as table
DateDaySessions

What verification found

8 “open” tasks checked against real code.

Half of what your plans call open is already finished.

✓ Resolved 4 ◐ Partial 3 ● Open 1
View as table
VerdictCount
Resolved4
Partial3
Genuinely open1

You keep building the hard 90% and stopping just before the cheap 10% that makes it land.

The pattern under four separate features

What you built

3 headline · 5 more · 6 fixed
The headline

Every AI in the app now advises through your own rules

Sit through the Contract interview once — objective function, framing rules, the things that demotivate you — and it flows into expert meetings, Mira, and fresh-project Foundation generation.

This is the moment the Dossier stopped being something you read and became something the app consults. It is also the most opinionated thing in the product: you built a system that takes instruction from you about how to speak to you.

Portfolio values synthesis

Reads five Foundation layers across every project and separates well-evidenced claims from thin ones — only the evidenced ones travel into exports.

Mockup Lab tells the truth now

Imported files become read-only references you fork. Before this, iterating never touched what Claude Code wrote.

Build a Package

Check off exactly the sections and projects you want instead of choosing between four fixed dumps.

/cr-review

Argue with the work card-by-card instead of forcing design opinions through a pass/fail QA frame.

Also fixed
  • 426 plans — every plan from the global directory had its path recorded twice, silently killing checkbox write-back
  • Expert prompts — three experts fell through to a generic default on any machine but yours
  • Note context — attached files and blueprints were dropped from planning prompts on three surfaces
  • The guided demo — was overwriting your real creative brief

What it revealed

6 patterns

Two of these come with a chart because the shape is the argument.

Notes completed, by the month they were captured

The scratchpad has capture and stages but no eviction. Anything not acted on within ~48 hours is effectively dead on arrival.
View as table
CohortEventually completed
February88
March18
April6
May1

You built four things that require you to remember them

Decision memory has zero callers on its own surfacing hook. Drift alerts only run on .onAppear. The Claude Code sync writes on enable and never after. Chronicle search stops at the archive boundary without saying so.
Act on this4 suggested tasks

Still open

7 confirmed · triaged

Every item here was checked against the code, not read off a checkbox — and then sorted by what to actually do about it.

Fix before anything else

Data integrity, and a guardrail that undercuts its own feature. Each has a matching task in Suggested.
Fix now~1 hour total
01
The cold archive can lose or hide 927 records
Two paths make today's intact data luck rather than design: a lost 101-byte sidecar hides the whole archive, and a decode failure lets the next run atomically overwrite it.
How I verified it
Hot 175 + cold 927 = 1102, zero duplicate ids, zero overlap, all 1072 pre-archival records addressable. Plan snapshots: 325 stubs ↔ 325 archived. Holes at ChronicleManager.swift:563 and :177.
Fix now
02
The three experts that exposed the wiring bug are still unwired
You shipped Craft Expert to stop experts being half-registered. Its three founding cases are in AgentCatalog but neither AgentInfo nor AgentModeMap.
Fix now

Finish the last wire

Three features that shipped at 90% — each is doing a fraction of its work because the surfacing step was left for later.
Complete~7 hours total
03
Drift alerts don't actually read your Contract
The baseline is hardcoded prose. watchOutFor appears nowhere in the detector — so the alerts can misquote your own stated goals back at you.
Complete
04
Decision memory can be searched but never speaks up
relatedDecisions exists, is documented as the surfacing hook, and has zero callers repo-wide.
Complete
05
VentureRaven can't read the profile you're exporting to it
Write side complete and versioned. Read side absent across all 1,389 Swift files in the fork.
Complete

Decide, don't build

These don't need code yet — they need a call from you about whether they're still worth it.
Needs a decision~15 min
06
Two ideation sessions still can't coexist — 41 tasks, none started
Live failures today, but you may not run concurrent sessions often enough to justify it. Commit or shelve; right now it inflates every open count.
Decide
07
Opening an old plan from an Objective shows a blank page
Filed under “accepted limitations,” which undersells it — compacted plans render empty, which reads as data loss.
Decide

Already done

nothing to do here

Recorded as open, verified as finished. These need no action beyond a store re-sync.

Verified complete

The Dossier's four biggest “remaining” tasks, and 16 of 17 Mockup Lab tasks

All shipped, wired, and confirmed by call site rather than by commit message. The plans simply never got updated.

  • Portfolio Pattern Report agent
  • DossierContextBuilder injection ×3
  • Values section + provenance
  • Claude Code projection auto-write
  • Mockup import + HTML round-trip
  • Opt-in CLI streaming + bubble

Suggested tasks

24 in 7 batches

Grouped so you can hand off a whole batch at once. Items paired to a Still Open card above are already counted there.

Make your records trustworthy again

Four things that each make the app say something untrue — about what's archived, what's done, and what a plan contains.
Fix now~1½ hours
Close the two archival holes
Code · ~30 min
Re-sync the two stale plan task stores
Code · ~20 min
Run the archival Phase 1 verification
Testing · ~15 min hands-on

Turn your best instinct into a system

You catch silent failures better than most people, one at a time, by hand. These give it somewhere to live.
Worth starting~5 hours
Author /cr-silent — a repeatable silent-failure sweep
Feature · ~4 hours
Promote the rule to a repo-wide convention in CLAUDE.md
Code · ~1 hour

…5 more batches (Design register · Make the app tell you · Backlogs that never drain · Outside eyes · Blueprints)

Loose threads

8 triaged

Close — events already answered these

19 notes across four groups. About 16% of the open backlog, none of it ever an intention.
Close~10 min, one pass
15 stale “Continue:” handoff notes
3–5 months idle · auto-generated resume payloads for work that already shipped
Close
Mockup round-trip spec — a phased plan for a solved problem
5.3 months idle · shipped twice over
Close
“Delete WidgetCTAButton.swift and WidgetStatusDot.swift”
4.7 months idle · neither file exists in the tree
Close

Promote — smaller than their age suggests

Each unblocks something you shipped this fortnight.
Start
Deep-link “Open in Mockup Lab” to the actual session
1 day idle · finishes the round-trip you just shipped
Start
Note file-context silently missing from planning prompts
12 days idle · archived without ever being fixed
Start

What this read

38 Chronicle records · 26 commits · 36 task stores · 745 plan snapshots · 118 notes · 50 blueprints.
  • Nothing truncated. 38 records, well under the 150 cap; no range clamp.
  • 8 tasks verified against code, not checkboxes. ~266 others were not — treat that count as unreliable until the stores re-sync.
  • Feature and design refinements are proposals from a forward-looking pass, not verified findings.