00
The verdict
CYCLE 07 · 2026-10-20The money is in three setup choices: sessions keep running after the save warning ($36.1K at 400K+ context), helpers inherit the most expensive model ($44.9K on Opus), and the always-loaded floor rides every reply. The production-ship gate was an honor system (130 self-unlocks with no ship word). All ten were fixed the same day.
Figures are the whole 51-day parser window. The deltas under them compare the 50 days since the 2026-08-30 audit against the 1 before it.
58.9B
Tokens billed
▲ 1018.1% per active day
1.2B a day now, 105.3M before
$111,826
Opus-equivalent value
▲ 842.1% per active day
$2,232 a day now, $237 before
3.73%
Go / continue tax
▼ 3.97pp of typed turns
3.72% now, 7.69% before
01
The full record
It starts on 2026-06-19 with four words: install node and pnpm. From there the record runs to 2026-10-20, 124 days, 5,394 sessions, $445,495 of Opus-equivalent value. Claude Code prunes its own transcripts after about thirty days, so the live parser can only ever see the most recent 51. The other 14 days survive here because the archive is append-only. Without it, this page would quietly forget where it came from.
02
The sheet
51 rows, one per day, nothing aggregatedEvery day in the current parser window, readable as a row rather than only as a shape. Amber marks the maximum in each column.
| Date | Day | Sessions | Opus-equiv | Tokens | Nudges | First prompt |
|---|
03
Contour survey
Sessions per day, banded by elevationEight contour bands, each an equal slice of the range, lit violet through teal as the elevation rises. Spot heights in amber mark every local peak.
04
The ribbon
Daily burn, lit by magnitudeDaily recordRibbon height = opus-equivalent burn
05
Orbit and signals
When it happens, and what to fixSession starts by local hour
Red = after 22:00
Signals
Worst first
06
The punchcard
Weekday against local hourWhere the two histograms cross. Each cell is one weekday-hour pair, sized by how many sessions started there.
07
The week
Session starts by weekdayThe punchcard collapsed onto one axis. Seven bars, one per weekday, counting every session that started on it.
08
The brief
Words in the day's opening prompt, averaged. A rising line is more front-loading and fewer follow-up nudges; a flat low line is a day spent steering instead of briefing.
09
The cache waterfall
96.7% of billed input is cache readAlmost nothing billed is new. 96.7% of every input token charged in this window is context read back out of cache, which is the same context paid for again on the next turn. The multiplier below is what one token of context actually costs across a session.
Billed input, 58.8B tokens
One bar, three buckets
Cache read 97.06% · 57.1BCache write 2.94% · 1.7BFresh input 0.00% · 562.9K
Every bucket at true scale
Cache read (re-billed context)57.1B
Cache write (context first paid for)1.7B
Output written by the model128.2M
Fresh input typed or read in562.9K
What the re-billing costs
96.7%
Cache hit ratio
cache read as a share of billed input
33.0x
Same ratio, this window
57.1B read / 1.7B written, 2026-08-29 to 2026-10-20
10
Model mix
By Opus-equivalent value, worst habit beside itValue by model
7 models fired
claude-opus-5-5$66,585
31.5B tokens · 59.5% of value
claude-opus-5$40,839
19.2B tokens · 36.5% of value
claude-sonnet-5$2,083
4B tokens · 1.9% of value
claude-sonnet-5-5$1,958
3.6B tokens · 1.8% of value
claude-fable-5-1$330
504.9M tokens · 0.3% of value
claude-haiku-4-5-20251001$22
87.2M tokens · 0.0% of value
claude-haiku-5-5$9
33.6M tokens · 0.0% of value
Opus on trivial work
A smaller model closes these
47
sessions
1.9% of the 2,498 sessions in the window. The parser counts a session as trivial when it ran on Opus with two or fewer typed turns, six or fewer tool calls and eight or fewer replies. No dollar figure is attached because the parser does not price this subset separately, and an invented one would not be a measurement.
11
The bill
The 15 costliest sessions on the recordRanked by Opus-equivalent value, not by length. Opening prompts are withheld on the public build.
| # | Day | Project | Tokens | Opus-equiv | |
|---|---|---|---|---|---|
| 01 | 2026-09-06 | macbook | 403.1M | $802 | |
| 02 | 2026-10-02 | macbook | 446.1M | $789 | |
| 03 | 2026-10-02 | macbook | 466.7M | $782 | |
| 04 | 2026-10-03 | macbook | 329.7M | $594 | |
| 05 | 2026-09-23 | macbook | 308.3M | $565 | |
| 06 | 2026-09-07 | macbook | 271.3M | $531 | |
| 07 | 2026-09-07 | macbook | 264.5M | $524 | |
| 08 | 2026-10-12 | macbook | 247.5M | $479 | |
| 09 | 2026-09-21 | macbook | 275.3M | $474 | |
| 10 | 2026-10-01 | macbook | 265.4M | $473 | |
| 11 | 2026-10-01 | macbook | 252.7M | $442 | |
| 12 | 2026-09-14 | macbook | 235.6M | $435 | |
| 13 | 2026-10-12 | macbook | 217.6M | $425 | |
| 14 | 2026-09-27 | macbook | 203.1M | $423 | |
| 15 | 2026-09-27 | macbook | 136M | $418 |
12
Parcel plan
Where the value went| Project | Parser key | Sessions | Tokens | Opus-equiv | Share |
|---|---|---|---|---|---|
| macbook | -Users-macbook | 1,301 | 57.6B | $108,871 | 97.36% |
| / | - | 697 | 450.4M | $1,504 | 1.35% |
| trading-brain | -Users-macbook-Claude-trading-brain | 153 | 104.5M | $459 | 0.41% |
| refinement-loop | -Users-macbook-Claude-xenia-brain-tools-refinement-loop | 45 | 151.2M | $428 | 0.38% |
| work | -Users-macbook-Claude-personal-brain-housing-rental-watch-work | 141 | 399.4M | $268 | 0.24% |
| mentor-v7 | -Users-macbook-Claude-mentor-v7 | 66 | 104.6M | $154 | 0.14% |
| hospitable-audit | -Users-macbook-Claude-xenia-brain-tools-hospitable-audit | 23 | 105M | $79 | 0.07% |
| work | -Users-macbook-Claude-personal-brain-shopping-canfield-watch-work | 17 | 25M | $17 | 0.02% |
| tmp | -private-tmp | 16 | 11.8M | $12 | 0.01% |
| talks-auto (worktree) | -Users-macbook-Claude-bodhicitta--claude-worktrees-talks-auto | 8 | 619K | $10 | 0.01% |
| scratchpad | -private-tmp-claude-501--Users-macbook-7ffe2cd2-4093-45ec-9913-bace55bf09b6-scratchpad | 9 | 1.7M | $9 | 0.01% |
| mentor-v72-01 (worktree) | -Users-macbook-Claude--claude-worktrees-mentor-v72-01 | 1 | 2.3M | $5 | 0.00% |
| hosp | -private-tmp-claude-501--Users-macbook-61e30cb1-59a5-48ce-a208-a6aa2d682bb3-scratchpad-hosp | 2 | 4.5M | $5 | 0.00% |
| drafttest | -private-tmp-claude-501--Users-macbook-2e642a6f-96b0-41c5-90e8-43a9aba6e433-scratchpad-drafttest | 1 | 79.8K | $1 | 0.00% |
| talk-0930 (worktree) | -Users-macbook-Claude-bodhicitta--claude-worktrees-talk-0930 | 1 | 79.8K | $1 | 0.00% |
| talk-pipeline (worktree) | -Users-macbook-Claude-bodhicitta--claude-worktrees-talk-pipeline | 1 | 74.6K | $1 | 0.00% |
| rental-watch | -Users-macbook-Claude-personal-brain-housing-rental-watch | 8 | 727.5K | $0 | 0.00% |
| hooktest | -private-tmp-claude-501--Users-macbook-1fba2599-6a83-433e-bb53-a28e18af3ab1-scratchpad-hooktest | 1 | 957.5K | $0 | 0.00% |
| probe-ups | -private-tmp-claude-501--Users-macbook-6f929f52-a8a7-4e73-b7ec-bd2810046d6d-scratchpad-probe-ups | 2 | 140.7K | $0 | 0.00% |
| Desktop | -Users-macbook-Desktop | 1 | 55.8K | $0 | 0.00% |
| hookprobe | -private-tmp-claude-501--Users-macbook-99a6aece-2d32-4181-9657-6afde147e482-scratchpad-hookprobe | 1 | 93.9K | $0 | 0.00% |
| market-brain | -Users-macbook-Claude-xenia-brain-market-brain | 3 | 0 | $0 | 0.00% |
13
The tool belt
The twenty most-called tools, at true scale against each other. One bar is the whole shape of the work.
14
The skill wall
One cell per installed skill. The dark cells are capability owned and never once reached for.
top 20 by sessions
fired at least once
never fired
The 10 of Jake's own that have never fired, by name (65 distinct skills did fire in this window, 26 of them from the 36 he wrote)
brandkitch-bookcompdailygrowthhigh-end-visual-designimagegen-frontend-webplumbingpricelabs-refreshsell
15
Repeated openers
40 openers typed three times or moreEvery opening prompt that started at least three sessions, 805 sessions between them. Anything on this list is a command that has not been written yet.
you are the mentor advisor autocapture step. do exactly what is listed here and nothing else. you a…170x
you are jake lee's rental housing watch, running headless on jake's mac every 3 hours. you have no…87x
you are the mentor advisor autocapture step. do exactly what is listed here and nothing else. you a…47x
you are the mentor v6.1 capture step. your entire authority is the mandate below. follow it exactly…42x
you are the mentor v6.1 capture step. your entire authority is the mandate below. follow it exactly…39x
you are the mentor advisor autocapture step. do exactly what is listed here and nothing else. you a…37x
you are the mentor v6.1 capture step. your entire authority is the mandate below. follow it exactly…34x
you are the mentor v6.1 capture step. your entire authority is the mandate below. follow it exactly…30x
you are jake lee's rental housing watch, running headless on jake's mac every 3 hours. you have no…27x
you are the mentor v6.1 capture step. your entire authority is the mandate below. follow it exactly…23x
you are the mentor advisor autocapture step. do exactly what is listed here and nothing else. you a…22x
you are the mentor v6.1 capture step. your entire authority is the mandate below. follow it exactly…22x
read-only task on the hospitable connector (sun mountain stays). call only these tools: get-alerts…18x
you are jake lee's canfield watch scanner, running headless on jake's mac once a day. you look for…16x
you are the mentor v6.1 capture step. your entire authority is the mandate below. follow it exactly…14x
you are the mentor v6.1 capture step. your entire authority is the mandate below. follow it exactly…12x
list the names of every tool you can call, one per line, and call none10x
you are the mentor v6.1 capture step. your entire authority is the mandate below. follow it exactly…10x
you are jake lee's rental housing watch, running headless on jake's mac every 3 hours. you have no…10x
you are the mentor v6.1 capture step. your entire authority is the mandate below. follow it exactly…9x
you are the scheduled mentor research desk run for the 11:45 et slot, in ~/claude/mentor-v7. do exa…9x
you are the scheduled mentor research desk run for the 10:45 et slot, in ~/claude/mentor-v7. do exa…9x
you are the mentor v6.1 capture step. your entire authority is the mandate below. follow it exactly…8x
you are the scheduled mentor research desk run for the 14:45 et slot, in ~/claude/mentor-v7. do exa…8x
you are the scheduled mentor research desk run for the 09:45 et slot, in ~/claude/mentor-v7. do exa…8x
you are the mentor v6.1 capture step. your entire authority is the mandate below. follow it exactly…7x
you are the mentor advisor autocapture step. do exactly what is listed here and nothing else. you a…7x
you are the mentor v6.1 capture step. your entire authority is the mandate below. follow it exactly…7x
you are the scheduled mentor research desk run for the 12:45 et slot, in ~/claude/mentor-v7. do exa…7x
you are the scheduled mentor research desk run for the 13:45 et slot, in ~/claude/mentor-v7. do exa…7x
you are jake lee's rental housing watch, running headless on jake's mac every 3 hours. you have no…7x
you are the mentor v6.1 capture step. your entire authority is the mandate below. follow it exactly…6x
read-only weekly rule check of the hospitable connector, sun mountain stays (20 properties). call o…5x
you are the mentor v6.1 capture step. your entire authority is the mandate below. follow it exactly…5x
you are the mentor daily alert refresh. your entire authority is the mandate below. follow it exact…5x
stop hook feedback: your reply used an em dash. claude.md §2a bans u+2014 everywhere including chat…5x
you are the mentor v6.1 capture step. your entire authority is the mandate below. follow it exactly…4x
you are the mentor v6.1 capture step. your entire authority is the mandate below. follow it exactly…4x
you are the mentor v6.1 capture step. your entire authority is the mandate below. follow it exactly…4x
you are the mentor v6.1 capture step. your entire authority is the mandate below. follow it exactly…4x
16
Maturity
Graded, each justified from a figureEfficiencyA
96.7% of 58.9B tokens were read from cache rather than rebuilt. That is the strongest number on this page. Points off only for the 47 sessions that opened the most expensive model for trivial work.
Skill masteryB
26 of 36 installed skills have fired at least once. Depth is real where it exists, but 10 skills have never been reached for.
AutomationB
Save and sync are habits now. Repo pulls and capture procedures are still hand-typed, and the repeated openers in the record show the same text retyped rather than parameterised.
Prompt qualityB+
Openers are specific and front-loaded, which is why cache performance holds at 96.7%. The drag is length variance: some days open at a few characters, others paste an entire procedure.
Front-loadingC+
468 go / continue messages, 3.73% of everything typed. The instruction that removes these is already written in this operator's own CLAUDE.md, which makes it a discipline gap rather than a knowledge gap.
17
The playbook
Ranked, each tied to a number above01
Streamline
Cut the go / continue tax
468 nudges across 12,554 messages, 3.73% of everything typed. Each one is a paid round trip that buys no new information.
Front-load the whole brief once, then let it run to done. The rule already exists in CLAUDE.md; the record says it is not being followed.
02
Downshift
Stop opening Opus for trivia
47 sessions look trivial by length and tool use, yet ran on the most expensive model available.
Route quick lookups, file reads and one-line edits to a smaller model. This is the cheapest correction on the page and it costs nothing in quality.
03
Adopt
Light up the dark half of the shelf
10 of 36 installed skills have never fired. That is capability already installed and already documented.
Pick three unused skills a week and force them into real work. The wall in section 08 is the checklist.
04
Shift
Move the late work earlier
After 22:00 the fault rate is 5.494 retries and errors per session against 4.509 in daylight, paid across 245 late sessions.
Cut the session when the fault rate turns. Work done past that point is being paid for twice.
05
Automate
Kill the hand-typed repo pull
The record shows the same openers retyped across sessions, including a full pasted capture procedure running to hundreds of characters.
One parameterised command per repeated opener. Each one removes the typing and the stray sessions that mistyping spawns.
18
The audit
Behavioral sweep, every 30 daysThe 30-day behavioral sweep. Where the instrument counts the tokens, the audit judges the habits and ships the fixes. Eight steps, one isolated context each, run every 30 days. These are the findings from the latest cycle and the baselines the next one has to beat.
Latest cycle2026-10-20 · 6 steps · 6 read-only subagents, report first; then Jake ordered every finding fixed the same day. Full report and fix table in history/cycle07/REPORT.md
The money is in three setup choices: sessions keep running after the save warning ($36.1K at 400K+ context), helpers inherit the most expensive model ($44.9K on Opus), and the always-loaded floor rides every reply. The production-ship gate was an honor system (130 self-unlocks with no ship word). All ten were fixed the same day.
Baselines to beat next cycle
32.5%, $36,147
Spend share at 400K+ context, main and helpers
share of bill · RED, flat at 46-50% every half-month; save-nudge gets the RESUME written then the session keeps going
$44,896, 40.4% of the bill
Helper spend on Opus
dollars · RED, about $35.9K above Sonnet prices
130 of 222
Production-ship unlocks with no ship word in the session
unlocks · RED, a safety gap rather than a cost
178 of 4,509
Self-stop nudges per typed turn
typed turns · AMBER, 51 came right after a Stop hook forced a reply
81.1% of tokens in 11.0% of sessions (300+ turns)
Session-length concentration
share · RED, unchanged since cycle 03
68 of 133 real runs (51.1%)
Sessions that ran a skill's own engine while never invoking the skill
pairs · AMBER, down from 93.6% but not because of the guard (16.5% follow-through) (was 73 of 78, 479 engine calls)
242 of 469
Real sessions that used cycle 05's one-call brain entry point
sessions · AMBER, up from 0 of 109, but sessions using it probe more (median 11 vs 8) and helpers use it 14.6% (was 0 of 109)
10 of 36
Own skills never fired
skills · AMBER, all never-used ones sit in the bare-name listing group
Top macro built, and runners-up
Prompt-gated ship token A UserPromptSubmit hook writes the ship token only when Jake's own prompt says ship, deploy or merge, and the PreToolUse gate refuses any token the session wrote. Closes the honor-system gap and drops a wasted call per deploy.Hard stop at the top context rung: refuse the next prompt and print the RESUME path plus the paste line.Default helper model Sonnet, Opus pinned on judgment agents only.Stop hook that refuses to end a turn while background work is running.Read guards and the skill-bypass guard quiet in helpers; bypass guard drops dead skill names.Audit corpus keeps only human-flagged records; guard-efficacy falls back to banked baselines.
The running record, every cycle, newest first
CYCLE 072026-10-20 · 6 steps · 16 items shipped
HOOKS6 shipped
Ship gate rebuilt: only Jake's own words (typed, clicked, or sent mid-turn) unlock it, a session-written token is refused, a bare go no longer counts. 31 of 31 phrase cases, 16 of 16 gate checks; proved live on this session.
context-gate.sh: past 550K tokens Jake's next message is held with the RESUME path and paste line; / commands and a + prefix pass. Held GO on a real 829K session.
subagent-checkpoint.sh: every helper writes partial results to disk after each step and uses brain.py first; a real helper confirmed receiving it.
claim-proof-guard.py: a fixed/live claim with nothing run after the change, or an unproven Safari open, goes back once. 1.3% of 600 real final turns.
Read guards main-session only; skill-bypass guard silent in helpers, quote-aware, audit instruments exempt; link guard narrowed (10 of 10); no-prose-tells blocks 90+ word paragraphs.
save-nudge and the save command no longer stop mid-task on a nudge-triggered save; the resume hook quotes cycle 07 numbers.
SETTINGS2 shipped
Helpers default to Sonnet (CLAUDE_CODE_SUBAGENT_MODEL); design agents inherit the main model.
New lean read-only reader agent for audit and research fan-outs.
SKILLS2 shipped
Audit corpus keeps only harness-marked human records (6,082 to 4,725); guard-efficacy falls back to banked baselines.
Mentor jobs skip for 90 minutes after 2 broker failures in a row; the desk waits for its background work; Refinement Loop pinned to Sonnet.
FOLLOW-UP (same night)6 shipped
Bare skill names traced to the listing allowance in the Claude Code binary (least-used skills lose descriptions first); SLASH_COMMAND_TOOL_CHAR_BUDGET=43000; fresh session: 0 own skills bare, was 16 to 19.
signals.py: the 6 unwatched cycle 07 baselines now re-measured daily, each reproduced within 2% on the cycle window; wired into the daily job.
watch.py: 8 new readers (6 signals, bare own skills, unreachable memory files); count metrics now per active day; 17 baselines banked.
Hook probes for the 29 hooks nothing tested; all 57 hooks healthy (35 guarding, 12 nudging, 9 runs, 1 unprobeable); WhatsApp probe matches the 10-05 rule.
Unreachable memory files: 0 of 1,084 (the 94 was a direct-links-only count); memory-reach.py keeps it at 0.
Route tests made hermetic; the 9-25 money-guard probes committed.
CYCLE 062026-08-30 · 8 steps · 24 items shipped
HOOKS9 shipped
content-guard-sweep.sh WIRED at Stop, closing the structural hole: the three content guards only ever watched Write and Edit while this machine is instructed to write through Bash. 9 of 9 tests green, 76ms, ratchet-only so it never blocks on a pre-existing backlog.
TWO FIRST-TOUCH HOOKS WIRED, which is this cycle's answer to its own biggest finding: brain-probe-nudge.sh fires once on the first shell read of the brain and names the one-call command, and skill-bypass-guard.sh speaks when a session shells into a skill engine with the owning skill never invoked. Both probe green and the bypass guard fired on a live command during the cycle. A false positive on a bare git status against an engine directory was fixed before wiring.
All three content hooks narrowed so the always-loaded surfaces are checked: no-em-dash, no-british-english and no-ai-copy-tells each exempted its own directory. 97 test cases now cover all three.
prompt-integrity-guard extended to the plain-stop case, which is 46.4% of bare nudges against the 20.0% it previously covered, with hold-language exclusions so it stays silent on a real section 4 GO. It fired correctly on Jake's own nudges during this cycle.
probe-run-guard retuned: it was capped at 2 fires per session, so avoidable probe waste actually ROSE from 3.70% to 8.00% after it was installed. Cap 2 to 12, threshold 200K to 120K, on arithmetic rather than caution.
save-nudge gained a repeating 550K rung; a session had run unwarned to 643K where a single turn costs about $1.20.
question-timing-guard defaulted to log-only: its premise inverted under replay. Sessions with 6+ ask rounds carry LESS rework (12.1% per typed turn) than 1 to 5 (16.0%), and "stop asking" appears zero times in 639 typed turns.
The orphan check over ~/.claude/hooks now returns nothing, so no mechanism built this cycle is sitting inert.
All three content hooks extended again to cover skill markdown, which loads in full every time a skill fires: 304 em dashes and 78 British spellings were sitting there unseen. 157 cleared from run-time skill docs, 99 test cases now cover the three hooks.
SKILLS7 shipped
build.py: every money and token surface now reconciles to the headline. The section titled The bill and the cache waterfall had never been converted, so the live page carried $831 and 13.9B under a $15,006 headline, and the two tables that were converted used the record ratio 1.896 where cost needs 1.986.
tests/test_build_reconciles.py: nothing in this repo tested build.py at all, which is how a page could ship contradicting itself and still pass. It fails on purpose against the previous build.py.
human-corpus.py fixed twice: it discarded all 172 typed slash commands and counted 19 machine-written resume openers as Jake's, so every per-turn denominator in cycles 04 and 05 was about 3% high.
price.py globbed one directory too shallow and missed all 71 subagent transcripts; it now prints a subagent bill and a price-by-position table.
floor.py and listing.py: the standing floor is now priced as rendered, with a recorded bare-skill set that the instrument refuses to price against silently when stale.
skill-collision-lint: a blanket escape hatch waved through 10% of pairs, and skill-description-guard ran it in an isolation where the collision half was unreachable. Both fixed, and ~/.claude/commands joined the routing namespace.
usage-lens and publish moved to directory-scoped skills inside their own repos.
BRAIN7 shipped
The belief namespace gate was wrong in both directions for the third time in two cycles: it accepted tools, docs and beliefs as brain tags while refusing lincoln, a real brain with a live email log. Now derived from the registry plus the tree, with a test that fails if any real brain maps to a refused tag.
brain.py: a rarity gate so an answer is pinned on the query's rarest term, filename ranking by the brain's own naming law, a credentials route where there was none, and market data no longer prefix-matching Palisade to Palisades Park.
PRIMER.md brought inside its own budget for the first time, 24,563 to 22,664 bytes, per-brain beliefs 4 to 5.
Audio and video answerability went 0 of 24 to 24 of 24, 39 PDF twins written, 53 intra-day briefings fenced so one fact stopped appearing in 9 files.
Fence counts and the per-call cost constant were quoted in five files and stale by a cycle; both now have one measured home.
The cognition loop got the three act-verbs it never had: verify (re-stamp a stale head, optionally at a LOWER confidence, which the store previously could not represent at all), rule (promote or reject a truth fork from the front door), and a windowed brier so calibration is read as a trend rather than a lifetime scar.
The brain-to-Ops seam was frozen and reading healthy: the push-out map was still keyed by pre-rename scopes, so it exported 60 of 221 heads while its log said 'imported 0, 88 already current', which is what a healthy no-op and a totally broken seam both look like. 161 heads recovered, 109 of them Sun Mountain.
CLAUDE.MD1 shipped
Replaced again at 24,682 bytes, preflighted clean through all three content hooks before applying. 9 numeric or premise claims corrected, 5 evidence claims removed as unreproducible, 6 blocks cut as prose standing in for a mechanism, 3 rules added that the cycle earned.
CYCLE 052026-08-30 · 8 steps · 20 items shipped
HOOKS6 shipped
save-nudge.sh reworked: the 400K rung now blocks and makes the model write the RESUME file itself, instead of a systemMessage addressed to Jake that the model never saw. 27 of the 45 sessions that crossed 400K had no resume file, and those 45 burned 82% of the bill.
Three guards built this cycle registered in settings.json and proved live: question-timing-guard (a 6th ask round in a session that front-loaded nothing), probe-run-guard (a third look-only shell call in a row on a 420K session), skill-description-guard (a SKILL.md whose description will not parse).
question-timing-guard now emits additionalContext on stdout rather than stderr, so it reaches the model rather than only the human.
grep-fence-guard: closed a leak where a recursive grep run after a cd resolved against the original directory, which is the shape behind 1,346 of 1,772 brain commands. Self-test extended to 21 cases, all pass, and a false positive on targeted greps fixed.
Five guards that had never been probed are now proven: monitor-notify-cost, no-british-english, no-prose-tells, prompt-integrity-guard, no-secrets-in-memory.
cleanupPeriodDays raised 7 to 45 so the next sweep reads a real window.
SKILLS6 shipped
/daily repaired: an undoubled apostrophe inside a single-quoted YAML scalar closed the string early, failed the whole frontmatter, and made the listing fall back to the body heading, so the skill was invisible to routing while Jake typed nine messages about his docket.
/market, /money and /ship descriptions rewritten against measured misses from Jake's own typed corpus.
visual-tuning-board.md: five options is now the default and never asked for, the trigger covers any single component about to be built, and every board must print the premise all options share as a sentence a person can disagree with.
New in the audit skill: substrate.sh (prints the real window before any step runs), human-corpus.py (separates Jake's ~624 typed turns from the ~900 raw user records), memory-load-check.py, price.py, skill-collision-lint frontmatter checks.
parse.py dedupes usage by requestId and emits the true series alongside the legacy one; the unused-skills instrument now requires a SKILL.md and counts own versus plugin skills apart, 112/86 to 36/21.
watch.py: probes added for every previously unproven guard, a reader for MEMORY.md overflow, and it now reads both stdout and stderr so a guard talking to the wrong reader cannot look healthy.
BRAIN7 shipped
tools/brain.py: new one-probe CLI, 5 of 5 audited questions answered in a single call against a 9-probe session median.
primer.py: per-brain grouping with confidence ranking and visible cut counts, 15 of 15 brains render against 1 of 15 before.
Belief tag namespace normalised 24 tags to 15 across four stores, 484 records, with a writer-side validator.
rankings.md now carries all 251 banked geographies, so Kingston, Saugerties and Ulster County resolve; the documented route had been answering a false no to a question Jake actually asked this window.
sun-mountain-stays INDEX.md split 43,906 to 2,580 bytes; proof-ledger routed from the master index after 85 touches with no index row anywhere.
Conflicts no longer expire, primers name every truth fork and every prediction due inside 7 days, and the scorecard now surfaces the 126 predictions the nightly sweep was binning unscored while reporting overdue: 0.
First repair ever issued against a belief that said something was broken: B-0573 stood at confidence 0.95 for 12 days after the code fix had already landed.
CLAUDE.MD1 shipped
Full proposed replacement written and held for review at ~/.claude/CLAUDE.md.audit05-proposed, with a per-change justification at CLAUDE.md.audit05-changes.md: 17 stale or wrong claims corrected, 3 new measured rules, 7 lines cut where a mechanism now exists.
CYCLE 042026-08-25 · 8 steps · 18 items shipped
HOOKS3 shipped
prompt-integrity-guard.sh: injects the resume protocol when a bare nudge follows a dead stream, and refuses to let a 50,000-char truncated paste be filed as a whole record. 17 of 17 tests, proved to fail on purpose, 100% recall and zero false positives on a full-corpus replay.
save-nudge.sh re-pointed from transcript line count to real context size: precision rises 63.3% to 78.4% at rung 1 and 98.4% at rung 2, cutting false fires from 168 to 80.
REJECTED with its number: a length-blocking hook (25-30% precision against an 18.2% base rate), a full-read warning (would have fired once in 30 days), and a narration-stall patch (17.9% precision, would block 64 legitimate stops to catch 14).
CLAUDE.MD5 shipped
Rewritten and 13.8% shorter: 28,594 to 24,636 bytes. Backup kept.
Deleted three measurably false claims that were billed in every session, including images at 49.4% of context.
Deleted the character caps: correction rate is flat near 5% from 500 to 4,000+ chars, and the caps sat below what Jake actually accepts.
Added the rule his own memory already held but the file never stated: a bare 'ship it' IS the GO through production. He had been re-authorizing by hand in 128 sessions.
Deleted the disproven single-quoted-YAML rule; a rule that looks actionable and fixes nothing is worse than no rule.
SKILLS4 shipped
Four descriptions rewritten for trigger accuracy, including one that routed work to a skill retired 2026-08-16.
The worst SKILL.md body rewritten (email, 655 lines, the only skill whose description contradicted its own body).
skill-collision-lint.py repaired: it had reported clean vacuously for five weeks by skipping any skill that owned its own name-word. Proved to fail on purpose.
Board mode wired into /perfect, /upgrade, /polish and /design through one shared spec, worth a measured 1.76 turns per visual episode.
BRAIN6 shipped
Long rows narrowed: lines over 1,000 characters fell from 434 to 20, so a single grep hit no longer costs 2,187 tokens.
301 PDF text twins written, so all 557 PDFs are now greppable; 12 genuinely need OCR.
Found and fixed the duplicate-writer root cause: a briefing filename keyed on the run timestamp, which twinned an unchanged digest on every re-render. 51 duplicates archived, three regression tests added.
Coverage bug fixed in the cognition scorecard itself: it was keyed on the human label, so scopes sharing a label silently overwrote each other. It was hiding 6 of 24 scopes and reported 26 stale against a true 48.
Four silent mechanisms rewired (/market, /money, /signal, /growth): the cognition step sat as an appendix after the run sequence, so nothing in the numbered procedure ever reached it.
Nothing deleted. Retired content moved to a fenced archive, and the two stale worktrees were fenced rather than removed because that ruling is Jake's.
CYCLE 032026-08-07 · 8 steps · 25 items shipped
HOOKS9 shipped
dup-read-guard rebuilt: keyed on transcript_path instead of session_id, so sibling subagents no longer share one dedup log; ranged reads no longer poison a path; a file edited since it was read now passes.
big-read-guard: closed the 43.0% limit escape hatch, added a 30,000-byte threshold alongside the 500-line one (calibrated to the measured 61.3 bytes/line), and fixed a latent bug where a path containing a space silently disabled the guard.
image-read-guard retuned 250KB to 150KB, covering 70.1% of image tokens instead of sitting above half the problem.
bash-cat-guard (new): closes the Bash route that bypassed all three Read guards, 1,842 measured calls. Narrow by design, so pipes, redirects and capped head/tail pass untouched.
grep-fence-guard (new): denies recursive plain grep inside the brain, because .ignore is ripgrep-only and 88.2% of searches were plain grep.
no-checkin-guard rebuilt and regression-tested on 762 real cases: recall 23.1% to 55.4%, and it no longer fires on API-error restarts.
big-read-guard: exempted images and PDFs from the new byte axis. It was judging a 64KB JPEG by bytes-per-line, which is meaningless for binary, and denying reads that image-read-guard already governs at 150KB. Found by the guard blocking this cycle's own screenshot review.
dup-read logging split from PreToolUse to a new PostToolUse companion (dup-read-log.sh). Both guards ran on Pre, so when big-read-guard denied an oversized read, dup-read-guard had already logged the path: nothing entered context, yet the retry was refused as a duplicate of a read that never happened. Pre denies, Post logs.
deliverable-link-guard (new, Stop): blocks a final message that tells Jake to go look at something while containing no URL, no absolute path and no markdown/backtick file link. Targets his #1 correction (95 messages, 69 sessions, 45% of triggering drafts had no link anywhere). 14 of 14 test cases correct, and an explicit 'not deployed' or HELD always passes.
CLAUDE.MD6 shipped
Reconciled from 401 lines back down to 319 lines / 24,584 chars, below the pre-cycle floor, with all four steps' edit markers removed.
Resolved the file's sharpest self-contradiction: ask-as-many-rounds-as-it-takes versus default-is-ship, now one rule with a narrow trigger.
Promoted the ship-the-link rule up beside answer-first, since the top correction in the corpus is an unreachable artifact.
Section 9 no longer inlines the ship Done-checklist, which was suppressing the /ship skill in 92 of 125 ship-request sessions.
Fixed a factually wrong rule: the booking engine is per-repo, not universally Hospitable.
Added one line only, on project-file precedence, backed by 8 measured contradictions across 16 project CLAUDE.md files.
SKILLS5 shipped
Restored 6 skills to the live listing by fixing malformed YAML descriptions (daily, sell, signal, rotate, publish, pricelabs-refresh).
Rewrote the 3 worst descriptions for trigger accuracy (frontend-design-plus, teardown, imagegen-frontend-web), each now scoring 5 of 5 on trigger simulation.
Rewrote imagegen-frontend-web's body from 987 lines to 169, cutting generic design advice the model already knows and adding the token cost gate it never mentioned.
Rehomed the retired /pulse duties: /briefing now owns closing out every open cognition prediction, not just its own.
Removed dead /pulse and /state routing from 6 skills that named two nonexistent surfaces as the reason to write to the loop.
BRAIN5 shipped
Retrieval maps and 5-line abstracts generated on all 82 qualifying docs (the gap was 78 files, not the 7 first measured).
6 missing directory indexes written: tools, docs, _briefings, comp-brain, _daily, beliefs.
Fenced the worktree shadow (6,846 files, 54% of the repo) plus hostgpo raw JSON worth 316,828 tokens sitting on the hot path.
The refinement loop can now record a WIN, which it structurally could not before, and its hardcoded 0.7 confidence is now judged per prediction.
Wired build-maps, build-series-indexes and the intake guard into the weekly run so the maps cannot drift back.
CYCLE 022026-07-19 · 8 steps · 15 items shipped
HOOKS3 shipped
no-ai-logo-redraw.sh: PreToolUse guard denies gemini-nano-banana redraws of favicon/logo/OG/wordmark (tested 6/6)
image-read-guard.sh: denies raw image reads over 250KB, steers to sips downscale or dimension check
both registered + validated in settings.json (backed up first)
CLAUDE.MD4 shipped
reconciled 227 down to 216 lines (folded the Step 2 §10 block inline, deleted the wrapper)
new rule: Scope is a ceiling, not a floor (40 corrections) into §1
new rule: Touch only what I named + diff-check (~60 regressions) into §8
§6 gained the image-read-guard + gemini base64 clause; fixed cache-read 96.6 to 96.3; stripped 8 stray em dashes
SKILLS4 shipped
fixed the /upgrade line that told Claude to GENERATE logos (the redraw-loop source; now uses the real repo mark per §8)
reconciled the /intake PII self-contradiction
rewrote airbnb / audit / ship descriptions with explicit NOT-triggers
24 skills audited, 0 dead, 0 deletes
BRAIN4 shipped
fenced market-brain/pricing/ in .ignore: -51.8% cold-grep bytes, -22.7% avg tokens/answer (revenue queries -39.8%)
extracted 54 PDF text-twins; stubbed 18 voice-memo + priority caption twins (Whisper/OCR queued: 21 + ~30 items)
head-read abstracts on the 5 files over 500 lines; INDEX + brain CLAUDE.md updated
cognition: all 7 participants verified correctly wired, no fix needed; surfaced a silent sync-brain-refresh prod-deploy failure
CYCLE 012026-07-15 · 7 steps · 13 items shipped
HOOKS3 shipped
no-checkin-guard.sh blocks banned progress check-ins
dup-read-guard.sh denies same-session re-reads
big-read-guard held at 500 lines
CLAUDE.MD3 shipped
reconciled 208 down to 197 lines
added the message-draft defaults block
§6 gained Edit-over-Write, images-priciest, cap-bash
SKILLS3 shipped
24 down to 21 (cut find-skills + industrial-brutalist-ui, merged design-taste)
trigger rewrites on signal / xenia / brandkit
sell routed through the market-brain budget ledger
BRAIN4 shipped
Columbus owner-fee backfilled (was a retrieval miss)
abstracts on 3 heavy indexes cut read cost 50-57%
39 call/image files renamed YYYY-MM-DD-entity-topic
committed on feat/money-v2