Caverio trader companion, overnight programme 2026-09-26, T8

T8: Wireframes v0, the logic of the interface

Worker T8, 2026-09-26, started 00:37 Beirut. Planning only. Nothing outside companion-2026-09-26/ written. No colour, no type, no brand visuals. Page and component names are working names.

What this is. The skeleton Thomas asked for: what you see where, in what form, for the three layers. Every component is a numbered slot; the tables, the HTML wireframes in wireframes/ and the ASCII sketches below use the same numbers. Open wireframes/index.html for the list of wireframes; each one has a desktop grid and a phone column (resize under 700 px, or open with a phone).

Earlier work is experimentation. The 09-17 mockups (D1 to D8) and the live signal-room pages were read for what they taught, not carried forward. Kept as lessons: one clock for price and events (D1/D2), numbers instead of adjectives ("liq 1.4% of cap", not "weak"), hatched "unknown" instead of zero, sources and clocks per line. Dropped: the ladder and its stages (Forming / Confirming / Fading), stage pills, the rung checklist, the People page as a separate level.

Sizing basis (ours, script scripts/T8-intake-volume.py, output data/T8-intake-volume.txt, grade A). 09-11 to 09-24: a median of 402 distinct tokens a day touch events.sqlite (max 518); by source the medians are onchain-watchtower 232, fomo-alerts 182, proof 128, social-watch 9, the five caller channels 1 to 7 each. The old state machine opened a median of ~55 new situations a day (range 24 to 95). So the intake is about 400 tokens a day and the useful set is about 50; the look list has to be the 50, ranked, not the 400.

Contents

  1. Page map and navigation
  2. Global frame (on every page)
  3. Market brief (Layer 1)
  4. Explore, the look list (Layer 2 intake) and the look-list logic
  5. Situation room (Layer 2), one skeleton, three band profiles
  6. Hold-stage view
  7. Strategy zones (Layer 3 overlay)
  8. Record
  9. My board (watching / in / out)
  10. Data asks for T2 and T4
  11. Open design questions for Thomas
  12. What I did not do

Column key for every component table: # slot; Component; Question it answers; Field(s); Form; Time views (N now, T today, W week, M month); Refresh; Missing / stale; P priority (P0 must ship, P1, P2). "Ours" means computed from our own tape or machinery; "derived" means computed from a licensed source and shown only as a derived number until T1 rules otherwise.


1. Page map and navigation

                  +------------------+
                  |  MARKET BRIEF    |  Layer 1: the environment
                  |  /brief          |
                  +--------+---------+
       band row / chain row|   (opens Explore pre-filtered)
                           v
+-----------+     +------------------+      follow       +------------------+
|  SEARCH   |---->|  EXPLORE         |------------------>|  MY BOARD        |
| (any page)|     |  /explore        |                   |  /board          |
+-----+-----+     |  the look list   |                   | watching/in/out  |
      |           +--------+---------+                   +--------+---------+
      |                    | row                                  | card
      |                    v                                      v
      |           +------------------------------------------------------+
      +---------->|  SITUATION ROOM   /t/<chain>/<address>               |
                  |  one skeleton, band profile picks order + windows    |
                  |  [ Zones ] tab = Layer 3 overlay on the chart        |
                  |  when the user is "in": hold-stage view on top       |
                  +--------------------------+---------------------------+
                                             | "our record on this token"
                                             v
                  +------------------------------------------------------+
                  |  RECORD   /record  (public summary, member detail)   |
                  |  what the system saw and did; caller scorecards;     |
                  |  closed situations frozen as they were               |
                  +------------------------------------------------------+
Page Layer Job in one line Entry points Exits
Market brief 1 Is the environment giving or taking; where is money going; which band is hot nav, first page of the day band row or chain row to Explore (filtered)
Explore 2 intake What is worth a look right now, ranked by how much has lit nav, brief rows, alert "new situation" row to Situation room; follow to My board
Situation room 2 Everything on one token on one clock, so buy / no-buy / size / exit is one screen Explore row, board card, search, alert, Record row follow state, Zones tab, Record
Zones (tab) 3 Where this token sits against a strategy's buy / hold / sell zones, and how that strategy has done in this band tab on Situation room back to the room
Record trust What the system saw and did, dated, with outcomes; caller scorecards nav, footer, "our record" slot Situation room (frozen, read-only for closed ones)
My board personal The tokens I care about, grouped watching / in / out, with what changed since I last looked nav, alert Situation room (hold view if in)

Search lives in the frame. An address or ticker that we already track opens its room. One we do not track opens a room in "not yet in our view" state (header, market block, safety, and a "start watching" control that asks the backend to open a situation on demand; data ask DA-24).

State that persists across pages (per user, saved server-side once accounts exist, localStorage before that):

State Set where Used where
Time view (now / today / week / month) frame F4 every page; each component maps it to its own windows
Chain filter, band filter frame F5, brief rows Explore, Record
Follow state per token (watching / in / out) and the date it was set room S1, Explore row, board board, Explore row marker, alerts
Entry for "in" (optional: entry cap or price, entry time) room H0 hold view H1, chart marker
Last-seen time per token set when the room is opened "changed since you looked" counts on Explore and board
Zones tab state and chosen strategy room S14 room chart
Alert rules board M4 alert delivery

What does not persist: scroll, expanded rows, sort (the default sort returns each visit so the list means the same thing to everyone).


2. Global frame (on every page)

# Component Question Field(s) Form Time Refresh Missing / stale P
F1 Nav Where am I Brief, Explore, My board, Record four text links, current one marked n/a static n/a P0
F2 Search Take me to a token address or ticker, chain text field with results list n/a on type "not in our view" result offers to open a room P0
F3 Coverage and clocks Can I trust what I see right now per chain: tape lag (s), price lag (s); per source: last event age; X state one line of short facts, e.g. "BSC tape 25 s · SOL tape 40 s · RH ok · callers 3 m · X dark since 18 Sep" N 30 s any source older than 3x its cadence is printed with its age; dark sources say "dark since " P0
F4 Time view At what speed am I trading now / today / week / month four-way switch all on click n/a P0
F5 Filters Which part of the market chain, band two selects, shown only on Explore and Record all on click n/a P0
F6 Alerts Did something happen on a token I follow count of unread alerts count, opens the board inbox N push "alerts off" if no channel is set P1

F3 is the frame's most important slot. The audit showed the engine acts exactly where the stream is worst (14 of 131 deep crossings on a lone print, DATA-AUDIT §1.1). A trader must see the clocks before the numbers.


3. Market brief (Layer 1)

Wireframe: wireframes/brief.html.

The brief answers four questions in order, top to bottom: is money coming in or leaving; where is it going by chain and platform; where in the cap structure; what are the trenches doing on our own tape. Each block carries today / week / month through F4. The trenches block is the only one that is ours; everything above it is derived from public aggregates (T3 names sources and T1 rights).

# Component Question Field(s) Form Time Refresh Missing / stale P
B1 State line In one sentence, what is the environment doing composed from B2, B5, B8, B9: stablecoin change, top chain by volume change, hottest band, trenches 2x count vs 7-day median one sentence of facts, no verdict words ("Stablecoin supply +$1.2B this week; Solana DEX volume +18% vs its 7-day average; the $1M-10M band has the highest volume to cap; 14 tokens under $10M printed a 2x on our tape today, 7-day median 9") T W M hourly any missing clause is dropped and the sentence says which source is missing P0
B2 Stablecoin firepower How much dry money is parked, and is it growing total stablecoin supply; change over window; split USDT / USDC / other big number, change beside it, sparkline 90 d under it T W M hourly stale over 24 h: number greyed with "as of " P0
B3 Money in or out Is money entering crypto or leaving total crypto cap change vs stablecoin supply change per period; spot ETF net flow per day (BTC, ETH, SOL where listed) paired bars per period (cap change, stable change); ETF bars per day, in and out either side of zero W M daily ETF data only on US trading days; weekend gap labelled, not zero P1
B4 Where the dry money sits On which chains and exchanges is firepower parked stablecoin supply by chain, 7 d change; stablecoins on exchanges horizontal bars per chain, change as a number at bar end W M daily exchange reserves marked "estimate, " P1
B5 Chain heat Which chain is hot per chain (Solana, BSC, Base, Ethereum, Robinhood Chain, others by volume): DEX volume in window, change vs prior window, new pools count, bridge net flow in window table, one row per chain, sparkline column for volume; row is a link to Explore filtered by chain T W M hourly new-pool count only where we run a tape; other rows print "not on our tape" in that cell P0
B6 Terminal activity How much degen activity is there volume through trading terminals (Photon, Axiom, GMGN, BullX, Trojan, Padre, others) summed; share by terminal big number with change, stacked bar by terminal T W M daily if only some terminals are measured, the total says "N of M terminals measured" P1
B7 BTC vs alts Is money rotating into alts or hiding in BTC BTC dominance and its change; ETH/BTC; SOL/BTC; an alt-season index if T3 picks one three small numbers with sparklines, one line each W M hourly n/a (cheap, always available) P0
B8 Cap band heat Which market-cap band is hottest, money in relative to size per band (top 10, 11-50, 51-200, 201-1000, $10M-50M, $1M-10M, under $1M): volume / cap in window, its change vs prior window, count of tokens +100% in window, count of tokens -50% in window table, one row per band, a bar for volume / cap, rank number 1 to 7 in the first column; row is a link to Explore filtered by band T W M hourly (top bands), 15 min (trenches bands) bands under $50M come from our tape and cover only taped chains; the row says which chains P0
B9 Trenches on our tape What the meme end is doing, measured by us per day: swap volume on watched pools; tokens that printed a 2x within 24 h, by band; share of volume in the top 5 tokens; new pools opened small multiples: 7 daily bars for each of three measures, today's bar marked T W M 15 min per chain; a chain whose tape stopped shows a gap labelled "tape stopped P0
B10 Leverage Is the market levered up aggregate funding rate, open interest change, liquidations in window three numbers with change T W hourly P2 source not yet chosen P2
B11 Where traders are talking Where attention is, by venue calls per day across tracked caller channels; Fomo alert count per day; X mention count three daily bar rows, 7 days T W M hourly X row reads "dark since 18 Sep" until restored P1
B12 Sources and clocks Where did each number come from per slot: source, as-of time, method link small table at the foot n/a with each slot n/a P0

Why this order. Thomas's words: first money in and out and firepower, then by chain and platform, then BTC vs alts and cap bands, with today / week / month. B9 is added because it is the one block nobody else has and the one the trenches trader cares about most; it also ties Layer 1 to Layer 2 (same tape).

Forms used and why. Big number only where one number is the answer (B2, B6). Tables where the question is "which one" (B5, B8): a ranked table is faster to read than a chart for "which chain / band is hot". Sparklines for direction, not for values. No gauge on this page: nothing here has a meaningful fixed scale.

ASCII (phone reading)

+--------------------------------------------------+
| F1 Brief Explore Board Record   F2 [search.....] |
| F3 BSC tape 25s · SOL 40s · RH ok · X dark 18Sep |
| F4 ( now | TODAY | week | month )                 |
+--------------------------------------------------+
| B1 Stablecoins +$1.2B this week; Solana DEX vol  |
|    +18% vs 7d avg; $1M-10M band hottest ...      |
+------------------------+-------------------------+
| B2 STABLECOINS         | B7 BTC dom  58.1% -0.4  |
|    $284.1B   +1.2B 7d  |    ETH/BTC  .036  ~~~   |
|    ~~~~~~~~~~ 90d      |    SOL/BTC  .0021 ~~~   |
+------------------------+-------------------------+
| B3 cap chg vs stable chg (bars) | ETF flow/day    |
+--------------------------------------------------+
| B5 CHAIN HEAT  vol 24h  chg   new pools  bridge  |
|    Solana      $3.1B   +18%   11,240     +$40M   |
|    BSC         $1.4B   -4%    3,120      -$8M    |
|    Base ...                                       |
+--------------------------------------------------+
| B8 BAND HEAT  # vol/cap  chg  +100%  -50%        |
|    $1M-10M    1  ####    +22%  14     31         |
|    under $1M  2  ###     +5%   40     220        |
|    ...                                            |
+--------------------------------------------------+
| B9 OUR TAPE  volume | 2x count by band | top5 %  |
|    7 bars each, today marked                     |
+--------------------------------------------------+
| B4 by chain | B6 terminals | B11 venues | B10 lev |
| B12 sources and clocks                           |
+--------------------------------------------------+

4. Explore, the look list (Layer 2 intake)

Wireframe: wireframes/explore.html.

A look list, not a buy list. It is curated in the sense that a token needs at least one qualifying event from a tracked source to appear; it is loose in the sense that one event is enough. Rows rise as more lights come on.

# Component Question Field(s) Form Time Refresh Missing / stale P
E1 Brief strip What is the environment doing (without leaving this page) B1 sentence one line, links to Brief T W M hourly hidden if B1 missing P1
E2 Filters and sort Narrow the list chain, band (F5); source (callers / X / wallets / traders); "only new since I looked"; sort: lit (default) / newest / fastest selects and a sort switch all on click n/a P0
E3 Counters How busy is it situations opened in window, rows with 3+ lights now, rows that lit a new light in the last 15 min three small numbers N T W 1 min n/a P1
E4 Just lit What changed in the last 15 minutes up to 5 rows that lit a new light in the last 15 min, each with the light that came on and its clock short list above the main list, each line " · traders +1 · 4 m ago" N 1 min empty: "nothing new lit in 15 min" P0
E5 The list What is worth a look, ranked per row: see the row spec below table, 25 rows by default N T W 1 min (rank), 70 s (market numbers) row market numbers stale over 5 min show their age in the cell P0
E6 Cooling What was lit and has gone quiet rows whose lights all decayed within the window, collapsed one collapsed section, count in its title T W 5 min n/a P2
E7 Coverage note What this list cannot see chains not taped, sources dark one line under the list N 5 min n/a P0

The row (E5), left to right

Cell Field Form Why
E5.1 follow marker empty / watching / in / out mark, click cycles watching on and off one click from list to board
E5.2 symbol, chain, pool age text; age in minutes under 2 h, hours under 3 d, then days age sets which band profile the room will use
E5.3 market cap now and band number, band in small text the size question first
E5.4 cap change since first seen signed percent and the first-seen clock ("+84% since 19:39") ties the move to when we saw it, not to a random 24 h
E5.5 lights eight fixed cells in fixed order: callers, X, wallets, traders, flow, holders, attention, narrative; each cell empty (not lit), filled with a count (lit), or hatched (cannot be seen on this chain); the most recent light is marked the row's reason for being where it is, readable without a legend after one visit
E5.6 last change the newest event in words with its clock ("Sugar called, 4 m ago", "2 tracked wallets bought, 11 m ago") what just happened
E5.7 flags safety flags as words, only when present ("LP unlocked", "dev sold 40%") the no-buy reason, never hidden inside a score
E5.8 since you looked count of new events since the user's last-seen time on this token; blank if never opened the board habit on the list

No composite score is printed. The rank is visible through the lights and the "last change" clock.

How items appear

  1. An event from a tracked source names a token (T4 defines "tracked": caller channels, solid X accounts, tracked on-chain wallets, Fomo-profiled traders). The token is resolved to chain and address; unresolved mentions go to a holding area and never appear on the list.
  2. The first qualifying event opens a situation and the row appears with one light on. The row carries the source that opened it (the "first seen by" clock).
  3. Further events from other categories light further cells. A second event in the same category raises that cell's count, not a new light.
  4. Lights decay: a light with no new event inside its category window turns off (proposed windows: callers 24 h, X 12 h, wallets 6 h, traders 6 h, flow 1 h, holders 6 h, attention 6 h, narrative 3 d; T5 owns the final numbers). A row whose lights are all off moves to Cooling (E6); after 72 h in Cooling the situation closes and goes to the Record.
  5. Every lit cell has at least one event id behind it (the ladder's evidence rule survives: no light without an event). Clicking a cell in the room shows the events.

Order

Default sort "lit": number of lights currently on, then the number that came on in the last 60 min, then the time of the most recent light. Rationale: a token lit by four independent categories is more worth a look than one lit four times by the same category, and a token lighting fast is more worth a look than one that lit yesterday. Ties are broken by recency, not by market numbers, so the list is not a price-momentum screen.

Alternative sorts: "newest" (first-seen time, for the first-hour trader) and "fastest" (lights on in the last 60 min). No sort by price change: that is what every screener already does.

How a row becomes a situation the user follows

Every row already is a situation (the system opened it). "Following" is the user's act: the follow marker on the row, or the follow control in the room. Watching puts it on My board and makes its alerts eligible; "in" switches the room to the hold-stage view. Unfollowing does not close the situation; only decay closes it.

How many rows is too many

Measured intake (our stores, 09-11 to 09-24): about 400 distinct tokens a day touch any source; the old state machine opened about 55 situations a day. With 24 h to 72 h lifetimes, 55 to 150 situations are open at any time. That is too many to scan.

ASCII (phone reading)

+--------------------------------------------------+
| F1..F4 frame                                     |
| E1 Stablecoins +$1.2B wk; $1M-10M band hottest > |
| E2 [chain v][band v][source v] sort: LIT|new|fast|
| E3 open 63 · 3+ lights 12 · lit in 15m 4         |
+--------------------------------------------------+
| E4 JUST LIT                                      |
|  CATALYST · traders +1 · 4m                      |
|  PONS · holders (early buyers holding) · 9m      |
+--------------------------------------------------+
| E5  * SYMBOL chain age | cap band | since seen    |
|       [cal][X][wal][trd][flw][hld][att][nar]     |
|       last change            flags   new-since   |
|  --------------------------------------------    |
|  o PONS  sol 3h | $2.4M 1-10M | +84% since 19:39 |
|       [3 ][//][2 ][1 ][on][  ][on][  ]           |
|       2 tracked wallets bought, 11m       · 5    |
|  --------------------------------------------    |
|  * AI  bsc 40m | $380K <1M | +12% since 21:02    |
|       [1 ][//][  ][2 ][on][  ][  ][  ]           |
|       Sugar called, 4m        LP unlocked        |
|  ... 25 rows ...                  [show more]    |
+--------------------------------------------------+
| E6 Cooling (18) >                                |
| E7 Not seen: Base tape, X accounts (dark 18 Sep) |
+--------------------------------------------------+
  [2 ] lit, count 2   [  ] not lit   [//] cannot see

5. Situation room (Layer 2), one skeleton, three band profiles

Wireframes: wireframes/situation-first-hour.html, wireframes/situation-day2.html, wireframes/situation-30m-narrative.html.

The answer to "does one skeleton serve all three": yes, with a band profile. The room always has the same fourteen slots. A band profile, chosen from pool age and market cap, sets three things: the order of the slots, the default time window of each, and which sub-panel of a slot is open. Two real forks exist, and they are sub-panels, not new slots:

Profiles (proposed; T5 sets the band edges from outcomes): first hour = pool age under 2 h, any cap; trenches = age 2 h to 7 d and cap under $10M; established = cap $10M+ or age over 7 d. The profile is printed in the header ("profile: first hour") so the trader knows why the page looks as it does, and can be overridden by a switch.

Skeleton (slot numbers are the default order)

# Slot Job
S1 Header and follow identity, address, age, band, first seen by whom and when, follow state
S2 Since you looked what changed since the user last opened this room (or in the last hour if never)
S3 Market block cap, price, liquidity, volume as big numbers with change
S4 Chart on one clock cap or price over time with every event as a marker on the same axis; Zones tab (Layer 3) lives here
S5 Lights the eight categories, each lit / not lit / cannot see, with count and first-lit clock
S6 Money flow net buy and sell pressure per interval, buyer and seller counts, size mix; S6a first-hour tape
S7 Holders early-buyer retention, concentration clock, who was already in before the first call, recurring wallets
S8 Attention and callers mentions by source over time, band-relative intensity, callers and their stance now, attention-to-money lag
S9 Tracked participants tracked wallets and Fomo traders in this token: when in, how big, still holding, their usual hold time
S10 Safety contract flags, LP state, deployer and its history, top holder share
S11 Narrative and peers narrative tag, narrative heat in the band, peer tokens in the same narrative and how they moved
S12 Timeline every event, in order, with source and clock; filters by category
S13 Our record on this token earlier situations on this token or deployer; what our machinery did (crossings, refusals, marks)
S14 Unknowns and clocks what cannot be seen here and why; freshness of every source feeding this room

5.1 Profile A: a token in its first hour

Example shape: Solana or BSC pool opened 25 minutes ago, cap $180K, one caller and two tracked wallets in, no narrative.

What changes: time runs in minutes; safety and the first-hour tape move up because the no-buy reasons (unlocked LP, deployer dump, bundled supply) show first; attention is "who called and how long before the money came"; narrative is collapsed.

Order on the page: S1, S2, S3, S10, S4, S6 (with S6a open), S5, S7, S8, S9, S12, S13, S14; S11 collapsed.

# Component Question Field(s) Form Time Refresh Missing / stale P
S1 Header and follow What is this and how old symbol, name, chain, address (copy), pool open time and age in minutes, band, first seen by (source, clock), profile label, follow control (watch / in / out) one line of text and controls; age is large ("25 min") N 1 min (age) address unresolved: room does not open (holding area) P0
S2 Since you looked What happened while I was away count and one-line list of events since last-seen (or last 15 min) up to 5 lines, each with clock N 30 s "nothing new since 21:04" P0
S3 Market block How big, how liquid, how fast cap now; change since pool open and since first seen; price; liquidity and liq / cap %; volume since open; swap count since open big number for cap, four small numbers under it N 30 s on taped chains, 70 s otherwise stale print: number shows its age; lone print under 0.75x of both neighbours is not shown (phantom rule) and the prior value stays with "1 print held back" P0
S10 Safety Is there a no-buy reason I can see now mint and freeze authority (Solana) or owner functions (EVM); LP locked / burned / unlocked with amount; deployer wallet: share held, sold so far, earlier tokens count; top-10 holder share list of flag words with numbers ("LP 100% burned", "deployer holds 3.1%, sold 0") N 1 min unknown flags printed as "not checked" with reason, never as clean P0
S4 Chart on one clock How did it move and what happened when cap per minute since pool open from our tape (or derived where no tape); markers: calls, tracked wallet buys and sells, Fomo trader buys, deployer actions, LP events line or candle chart, minute bars, markers on the same x axis, strip of markers beneath (D1/D2 lesson) N (since open) 30 s chains without our tape: chart from derived source only if T1 allows, else markers-only strip with the market block; gaps drawn as gaps P0
S6 Money flow, with S6a open Who is buying, how many, how big net buy USD per minute; unique buyers and sellers per minute; buy size mix (under $100, $100-1K, $1K-10K, over $10K); S6a: share of supply bought in the first 3 blocks (sniper share), buys from wallets funded by one source in the same block (bundle share), deployer buys and sells bars per minute above and below zero; two small counts; a 4-part bar for size mix; S6a as three numbers with a sentence ("22% of supply bought in the first 3 blocks by 9 wallets") N 30 s not taped chain: whole slot says "no tape on " (hatched); S6a unknown until the tape covers pool open P0
S5 Lights Which categories are lit eight cells as on Explore, each with count, first-lit clock, last event clock; safety shown as a separate flag count row of eight labelled cells with numbers N 30 s hatched cells for sources that cannot see this chain P0
S7 Holders Are the first buyers staying; is supply concentrating first 20 buyers: holding / reduced / out counts; top-20 buyer share of supply over time since open (concentration clock); buyers before the first call vs after three-part bar for retention; line for concentration next to the price line on the same clock; two counts N 1 min holdings not covered on this chain: tape-derived net position per wallet (sum of buys minus sells since open) labelled "from tape, transfers not seen" P0
S8 Attention and callers Who is talking and did money follow callers who called it (channel, clock, cap at call, stance now); minutes from first call to first tape buy after it; mentions per 5 min by source small table for callers; one number for lag ("money followed in 3 min"); bars for mentions N 1 min X row "dark since 18 Sep" P0
S9 Tracked participants Is anyone I would follow in it tracked wallets and Fomo traders: name or short address, first buy clock, size, holding / reduced / out, their median hold time and how often they enter this early table, one row per participant, max 10 then "and N more" N 1 min holding state "not checked" on uncovered chains P1
S12 Timeline In what order did it happen every event: clock, source, kind, actor, size, link list, newest first, filter chips by category N 30 s n/a P0
S13 Our record Has our machinery seen this deployer or pattern before earlier situations on the same deployer; crossings and refusals on this token with reason short list N 5 min "first time we see this deployer" P1
S14 Unknowns and clocks What can I not see here per source: last event age; per missing angle: why (no tape, no holdings, X dark) list N 30 s n/a P0
S11 Narrative and peers (collapsed) name / ticker heuristic tag only one line n/a 5 min "not tagged yet" P2

5.2 Profile B: day 2, cap $1M-10M

Example shape: BSC token, 30 h old, cap $3.4M, three callers over the day, two tracked wallets holding, a small narrative tag ("AI agent").

What changes: time runs in 15 minutes and hours; the question is "are the early holders still here and is money still coming"; retention and flow move up; attention is now relative (is this token getting more or less attention than its band peers); narrative opens; first-hour tape is history.

Order on the page: S1, S2, S3, S4, S5, S7, S6, S8, S9, S11, S10, S12, S13, S14.

# Component Question Field(s) Form Time Refresh Missing / stale P
S1 Header and follow What is this as profile A; age in hours one line N 1 min as A P0
S2 Since you looked What changed as A; default window "since last visit" or last 6 h up to 5 lines N 30 s as A P0
S3 Market block How big, how liquid, where is it vs its range cap; change 1 h / 24 h / since first seen; price; liquidity and liq / cap; volume 24 h and volume / cap; distance from the high since first seen ("-32% from the 24 h high") big number + five small ones N T 70 s as A P0
S4 Chart on one clock How did it move and what happened when cap per 15 min since first seen (or since open), markers as A plus caller stance changes; toggle 1 h / 24 h / since seen; Zones tab chart with markers and marker strip T W 1 min as A P0
S5 Lights Which categories are lit, and which are fading eight cells, each with count and age of last event; cells within 25% of their decay window marked "fading" row of cells N 1 min as A P0
S7 Holders Are early holders still in first 50 buyers: holding / reduced / out counts and the share of supply each group holds; concentration clock over 48 h; holder count trend; recurring wallets (seen in earlier calls from the same caller) count 3-part bar with supply share; line; number with change; count with a sentence T 5 min as A; holder count needs provider (DA-08) P0
S6 Money flow Is money still arriving net buy USD per hour for 48 h; buyers and sellers per hour; size mix per day; largest 5 buys and sells in window bars per hour, counts, 4-part bar, small table T W 5 min as A P0
S8 Attention and callers Is attention rising or falling, relative to its peers mentions per hour by source (callers, Telegram chats, X, Fomo); rank of this token's mentions / hour among open situations in its band ("4th of 37 in $1M-10M"); callers table with stance now and every stance change with clock; attention-to-money lag per call bars per hour; rank as number with peer median bar; table; lag list T W 5 min stance "not read" where only the formatted call is captured (DA-03) P0
S9 Tracked participants Are the people I would follow still in as A plus P&L-free facts: entry clock, size, current state, days held vs their median table T 5 min as A P0
S11 Narrative and peers What story is this part of, and is that story hot narrative tag(s) and source of the tag; tokens with the same tag opened this week and their median move since first seen; this token's rank among them tag words; small table of peers (max 5); one number W 1 h "not tagged" with the tagging sources tried P1
S10 Safety Anything that changed on safety as A plus changes since day 1 (LP moved, deployer sold, authority changed) flag list, changes marked with clock T 5 min as A P0
S12 Timeline In what order as A; first hour collapsed into one block ("first hour: 212 swaps, 9 snipers, deployer sold 0") list T W 1 min n/a P0
S13 Our record What our machinery did as A plus our marks on this token at 1 h / 4 h / 24 h from first seen short list T 5 min as A P1
S14 Unknowns and clocks What can I not see as A list N 1 min n/a P0

5.3 Profile C: $30M token with a narrative

Example shape: Solana token, 9 days old, cap $30M, narrative "AI agents", covered by several callers and many X accounts, holder base in the thousands.

What changes: time runs in hours and days; the first-hour tape and per-wallet tables matter less; the holder base, narrative heat and relative attention matter more; slower traders use week and month views. Safety collapses to a line unless something changed.

Order on the page: S1, S2, S3, S11, S4, S8, S7, S6, S5, S9, S10, S12, S13, S14.

# Component Question Field(s) Form Time Refresh Missing / stale P
S1 Header and follow What is this as A; age in days; narrative tag next to the symbol one line N 5 min as A P0
S2 Since you looked What changed as A; default window "since last visit" or last 24 h up to 5 lines N 1 min as A P0
S3 Market block How big, where in its range cap; change 24 h / 7 d / 30 d; price; liquidity and liq / cap; volume 24 h and volume / cap vs band median; distance from all-time high; CEX listings count (P2) big number + small ones T W M 70 s as A P0
S11 Narrative and peers Is the story it belongs to getting money and attention narrative tag(s); narrative heat: summed volume / cap and mentions of all tokens with the tag, this week vs last; this token's share of the narrative's volume; top 5 peers with cap and 7 d move two numbers with change; share as a number; peer table W M 1 h tag uncertain: "tag from ticker only" printed P0
S4 Chart on one clock How did it move and what happened when cap per hour (7 d) or per 4 h (30 d); markers limited to calls, stance flips, large tracked buys and sells, narrative events (peer launches over $10M); Zones tab chart with markers W M 5 min as A P0
S8 Attention and callers Is attention rising faster or slower than the price mentions per day by source; rank among band peers and among narrative peers; attention vs volume ratio over time (rising attention with falling volume is the "exit under the noise" pattern); callers table with stance, flips highlighted bars per day; two ranks; a two-line chart (mentions, volume) on one clock; table W M 1 h as B P0
S7 Holders Is the holder base growing or distributing holder count and 7 d change; top-10 and top-50 share and their 7 d change; early-buyer retention (first 100) still shown but lower; wallets over $10K count trend numbers with change; small line for top-10 share W M 1 h holder count needs provider (DA-08); Solana holdings not covered (DA-06) P0
S6 Money flow Net buying or selling by day net buy USD per day (30 d); buyers vs sellers per day; large trades (over $10K) count per day bars per day, two counts W M 1 h as A P0
S5 Lights Which categories are lit as B; at this size most lights are usually on, so the cells show counts per day rather than on/off row of cells with counts T W 5 min as A P1
S9 Tracked participants Are tracked people accumulating or leaving counts only: tracked holders now vs 7 d ago; entries and exits this week; top 5 by size two numbers, small table W 1 h as A P1
S10 Safety Anything changed one line unless a change happened in window, then the flag list line W 1 h as A P1
S12 Timeline In what order filtered by default to calls, stance flips, large trades, narrative events; full list behind a switch list W M 5 min n/a P1
S13 Our record What our machinery did as B; plus the strategy zones record for this band (links to Zones tab) short list W M 1 h as A P1
S14 Unknowns and clocks What can I not see as A list N 5 min n/a P0

5.4 Where the three profiles differ, in one table

Slot First hour Day 2, $1M-10M $30M, narrative
Position of Safety 4th 11th 11th, one line
Time unit minute 15 min / hour hour / day
S6a first-hour tape open collapsed into S12 gone
S7 holders lead field concentration clock first-50 retention holder count trend, top-10 share
S8 attention lead field lag from call to money rank among band peers, stance changes attention vs volume, narrative rank
S11 narrative collapsed open, 10th open, 4th
S9 participants full table full table counts only

ASCII, profile B (day 2), phone reading

+--------------------------------------------------+
| S1 PONS · bsc · 30h · $1M-10M · seen 19:39 by    |
|    Sugar · profile: trenches  [watch|in|out]     |
+--------------------------------------------------+
| S2 SINCE YOU LOOKED (21:04)                      |
|    22:10 Columbus called · 22:31 2 early buyers  |
|    sold · 23:05 net flow out 3rd hour running    |
+------------------------+-------------------------+
| S3 $3.4M               | S4 CHART (cap, 15m)      |
|   +84% since seen      |   ___/\__/\___           |
|   -32% from 24h high   |  ^c  ^w ^c   ^s  (marks) |
|   liq $410K 12% of cap |  [1h|24h|since] [Zones]  |
|   vol 24h $5.1M 1.5x   |                          |
+------------------------+-------------------------+
| S5 LIGHTS [cal 3][X //][wal 2][trd 1][flw  ]     |
|           [hld 1][att on][nar 1]                 |
+--------------------------------------------------+
| S7 HOLDERS first 50: 31 hold | 12 reduced | 7 out|
|    they hold 18% of supply · top-20 share ~~~    |
|    before 1st call: 9 wallets · recurring: 3     |
+--------------------------------------------------+
| S6 FLOW net/hour 48h  ||||.||'''|'''  buyers 214 |
+--------------------------------------------------+
| S8 ATTENTION 4th of 37 in band · callers:        |
|    Sugar  19:39 $1.8M  bullish                   |
|    Pows   20:02 $2.2M  bullish -> bearish 23:40  |
+--------------------------------------------------+
| S9 participants | S11 narrative | S10 safety     |
| S12 timeline | S13 our record | S14 unknowns     |
+--------------------------------------------------+

ASCII, profile A (first hour), phone reading

+--------------------------------------------------+
| S1 AI · bsc · 25 min · <$1M · seen 21:02 by      |
|    2 tracked wallets · profile: first hour       |
+--------------------------------------------------+
| S2 since 21:10: Sugar called 21:14 · 3 buys >$1K |
+------------------------+-------------------------+
| S3 $180K               | S10 SAFETY              |
|   +140% since open     |  LP 100% burned         |
|   liq $61K 34% of cap  |  deployer holds 3.1%,   |
|   412 swaps since open |  sold 0 · top-10 41%    |
+------------------------+-------------------------+
| S4 CHART cap per minute since open, marks below  |
|    _/'-/''\_/'                                   |
|    ^w ^w  ^c                                     |
+--------------------------------------------------+
| S6 + S6a FLOW per minute |||'|||.|  buyers 96    |
|   22% of supply bought in first 3 blocks, 9 wallets|
|   bundle share 6% · deployer: no sells           |
+--------------------------------------------------+
| S5 LIGHTS [cal 1][X //][wal 2][trd  ][flw on]... |
| S7 top-20 share 38% -> 31% (line) · first 20: 17 |
|    hold · before 1st call: 12 wallets            |
| S8 Sugar 21:14 at $120K · money followed in 1 min|
| S9 participants | S12 timeline | S13 | S14       |
| S11 narrative: not tagged yet (collapsed)        |
+--------------------------------------------------+

ASCII, profile C ($30M, narrative), phone reading

+--------------------------------------------------+
| S1 AGENTX · sol · 9 d · $10M-50M · AI agents     |
|    profile: established  [watch|in|out]          |
+--------------------------------------------------+
| S2 since yesterday: 2 callers flipped bearish;   |
|    holder count +4% 24h                          |
+------------------------+-------------------------+
| S3 $30.2M              | S11 NARRATIVE AI agents |
|   -8% 24h +61% 7d      |  heat: vol/cap 0.9 (+30%|
|   -41% from ATH        |  wk/wk) · this token 22%|
|   vol/cap 0.4, band    |  of narrative volume    |
|   median 0.3           |  peers: A $44M +12% ... |
+------------------------+-------------------------+
| S4 CHART hourly 7d / 4h 30d, filtered markers    |
+--------------------------------------------------+
| S8 mentions/day ||||'| · 6th of 52 in band ·    |
|    mentions vs volume (two lines, one clock)     |
|    callers: 9, 2 flipped bearish since Tue       |
+--------------------------------------------------+
| S7 holders 8,410 +4% 7d · top-10 19% (-2 pts)    |
| S6 net flow per day 30d | S5 lights counts/day   |
| S9 tracked holders 14 (was 11) | S10 no changes  |
| S12 timeline (filtered) | S13 | S14              |
+--------------------------------------------------+

The HTML wireframes carry the full layout for all three profiles.


6. Hold-stage view

Wireframe: wireframes/situation-hold.html.

When the user marks a token "in" (S1 follow control), the same room gains a hold block on top and reorders. It is the same skeleton; nothing is removed. The trader manages the trade; the room only has to make it obvious when the reasons they entered are going away.

H0, marking "in". One tap. Optional fields: entry cap or entry price, entry time (default now). No size, no wallet link in v0 (see open question Q3). If no entry is given, the hold block measures from the moment "in" was tapped.

What moves to the top

Order when in: H1, H2, S2, S3, S4 (with entry marker), S7, S8, S6, S9, S5, then the rest as the profile orders them.

# Component Question Field(s) Form Time Refresh Missing / stale P
H1 Position line Where am I against my entry cap at entry vs now; change since entry; time held; change from the highest cap since entry ("-18% from the high since you entered") one line of numbers, the change since entry large N 30 s no entry cap given: measured from the "in" tap, and says so P0
H2 What changed since you got in Are the reasons I entered still true ordered list of hold events since entry, each a fact with clock; see the list below list, newest first, max 8 lines, "and N more" N 30 s "nothing on the watch list has happened since 21:40" P0
H3 Alerts on this token What will reach me when I am not looking the hold alert rules that apply and whether each is on list of switches n/a on change "alerts off: no channel set" with a link to settings P1
S4 Chart as the profile, plus an entry marker and a shaded band from entry to now P0
S7 Holders as the profile, but counts are "since your entry" by default P0
S8 Attention and callers as the profile, stance flips since entry at the top of the callers table P0

The H2 hold events (what a trader in the position needs to know first)

Priority order, highest first. Each line is a fact with a clock and a number, never a verdict.

Order Event Rule (proposed; T5 tunes) Example line Data
1 Liquidity pulled LP removed or liquidity down 20%+ within 15 min "Liquidity $410K to $260K in 9 min (LP removal 23:12)" tape LP events, DA-11
2 Deployer sold any deployer or deployer-funded wallet sell "Deployer sold 1.2% of supply, 23:05" tape + deployer id, DA-10
3 Early-holder exodus of the first 50 buyers, 5+ reduced or out since entry, or holders representing 5%+ of supply sold within 60 min "7 of the first 50 buyers sold since you entered, 4 in the last 30 min; they held 6.1% of supply" holdings / tape, DA-05, DA-06
4 Caller turned any caller that called this token posts a bearish or exit message about it "Pows: bullish 20:02 -> bearish 23:40 ('took profits, careful here')" T4 stance, DA-02, DA-03
5 Tracked participant out a tracked wallet or Fomo trader in this token reduces 50%+ "wizardofsoho sold 70% of position, 23:31; held 3 h 10 m (usual 9 h)" holdings / fomo, DA-07
6 Money turned net flow negative for 3 consecutive intervals of the profile's unit "Net flow out 3 hours running: -$42K, -$18K, -$61K" tape
7 Attention fading while price holds band rank for mentions falls 50%+ from its peak while cap is within 10% of its high "Mentions rank 4th to 19th of 37 since 20:00; cap -6% in the same time" T4 counts, DA-04
8 Price level user-set price or cap alerts only; the system does not propose levels "Cap below $2.5M (your alert)" market block

How a bearish caller flip is shown

How a holder exodus is shown

Alerts that exist in v0

Delivered to the channel the user sets on the board (M4). Watching tokens get alerts 1, 2, 4 and "new light lit". In tokens get all eight. Each alert is one message: token, event line as in H2, clock, link to the room. Quiet hours and a per-token mute exist. No alert says buy or sell.

ASCII, hold view (phone)

+--------------------------------------------------+
| S1 PONS · bsc · IN since 21:40  [watch|IN|out]   |
+--------------------------------------------------+
| H1 entry $2.9M -> now $3.4M  +17%  held 1h52m     |
|    -18% from the high since entry ($4.15M 22:20)  |
+--------------------------------------------------+
| H2 SINCE YOU GOT IN                               |
|  23:40 Pows: bullish 20:02 -> bearish 23:40       |
|  23:31 wizardofsoho sold 70% (held 3h10, usual 9h)|
|  23:20 7 of first 50 buyers sold since entry,     |
|        4 in last 30m, held 6.1% of supply         |
|  23:05 net flow out 3 hours running               |
+--------------------------------------------------+
| S3 market  | S4 chart with entry marker + band    |
| S7 retention: at entry 38|9|3   now 31|12|7       |
| S8 callers (flips on top) | S6 flow | S9 | S5 ... |
| H3 alerts on this token: [x]LP [x]deployer ...    |
+--------------------------------------------------+

7. Strategy zones (Layer 3 overlay)

Wireframe: wireframes/zones-overlay.html.

Zones are a tab on the room's chart (S4), not a product and not a page. Off by default. When on, the chart is shaded into buy / hold / sell zones for one chosen strategy, and a strip under the chart says how that strategy has done in this band, with the forgiving win defined in plain words.

# Component Question Field(s) Form Time Refresh Missing / stale P
Z1 Strategy picker Which way of trading am I looking through strategy name (T6 defines the set, e.g. "first call follow", "early-holder hold", "day-2 retest"); its one-line rule in plain words select + one sentence n/a on change "no strategy defined for this band" P1
Z2 Zone shading Where on this token's path would the strategy have been buying, holding, selling zone start and end times per state, from the strategy's trigger rules applied to this token shaded bands behind the S4 line, labelled buy / hold / sell as S4 as S4 zones need the same inputs as the strategy; missing input means no shading and a line saying which input P1
Z3 Where it is now Which zone is the token in right now current zone and since when one word and a clock under the chart ("hold zone since 22:05") N as S4 as Z2 P1
Z4 Strategy record in this band Has this strategy worked on tokens like this n triggers in band over the last 30 d; forgiving-win share (moved up at least X% within Y h and stayed there Z min); median time to the win threshold; share that lost 50% first four numbers with n printed; a line stating the win definition M daily n under 30: printed "too few to read" instead of a share P1
Z5 This token's previous triggers Did this strategy trigger here before, and what happened earlier trigger clocks on this token and the outcome at the win horizon small list W M 1 h none: "first trigger" P2

Why an overlay: the mandate says Layer 3 is later and harder, trade management stays with the trader, and the situation room must carry the hold stage. Putting zones on the same chart means they cannot drift from the evidence; putting them on a separate page would make a signals product.

ASCII

+--------------------------------------------------+
| S4 [Chart] [ZONES]  Z1 strategy: first call follow|
|    "buy within 30 min of a 2nd caller if flow in" |
|  |buy |     hold          | sell |               |
|  |////|___/\____/\___    |\\\\\\|  ...           |
| Z3 hold zone since 22:05                          |
| Z4 in $1M-10M, 30 d: n=64 · forgiving win 41% ·   |
|    median 2h10 to +40% · 22% hit -50% first       |
|    win = +40% within 24 h, held 30 min            |
+--------------------------------------------------+

The numbers in the sketch are placeholders, not measurements; T6 produces the real ones.


8. Record

Wireframe: wireframes/record.html.

The public record of what the system saw and did. Two audiences: prospects (is this real) and members (how did the sources do). Public version shows R1 to R4 summaries; members see the tables in full.

# Component Question Field(s) Form Time Refresh Missing / stale P
R1 What we saw How much did the system look at per day: tokens seen, situations opened, situations that lit 3+ lights, closed four daily bar rows W M daily days with a source outage are marked with the source name P0
R2 Caller scorecards How did each caller's calls do per channel: calls in window, median move at 15 m / 1 h / 4 h / 24 h from call price, share that reached +50% within 24 h, share that fell 50% first, median lag call to our first sight; censoring caveat ("tokens that died before 24 h are counted at their last price") table, one row per caller, sortable W M daily calls with no price at call are counted and listed separately, never dropped P0
R3 Situations archive What happened to what we showed per closed situation: token, opened, lights at open and at peak, cap at open, best cap within 24 h, cap at 24 h and 72 h; link to the frozen room table with filters (chain, band, lights at open) W M hourly market marks priced on honest fills, repeats and phantoms excluded (DATA-AUDIT traps) P0
R4 By lights at open Did more lights mean better outcomes for situations grouped by lights lit at open (1, 2, 3, 4+): n, median best move within 24 h, share +50%, share -50% small table M daily groups under 30 print "too few" P1
R5 Quiet hours When is each chain dead per chain: hour-of-day by day-of-week of new situations or tape volume heatmap (justified: two cyclic dimensions, one value) M daily stopped tapes shown as gaps P2
R6 Strategy zones record How each strategy has done by band Z4 fields for each strategy and band table M daily as Z4 P2
R7 Method and data health Can I trust these numbers phantom rate and stale-repeat rate per source per day; outages; definitions; scripts link short text and a table W M daily n/a P0

Frozen rooms: when a situation closes, its room is saved as it looked at close (all slots with their data) and served read-only from the Record. This is the "what was knowable at each point in time" moat from the brand brief, made visible.

ASCII

+--------------------------------------------------+
| R1 SEEN per day: tokens | opened | 3+ lights | closed
|    |||||||  |||||||  |||||||  |||||||             |
+--------------------------------------------------+
| R2 CALLERS  calls med15m med1h med24h +50% -50% lag
|    Sugar    41    +4%   +9%   -22%  29%  44%  5m |
|    Pows     88    ...                            |
|    * tokens that died before 24h counted at last |
+--------------------------------------------------+
| R3 ARCHIVE [chain][band][lights]                 |
|    PONS 09-25 19:39 lights 1->5 $1.8M best $4.1M |
+--------------------------------------------------+
| R4 by lights at open | R5 quiet hours | R7 method|
+--------------------------------------------------+

9. My board (watching / in / out)

Wireframe: wireframes/board.html.

# Component Question Field(s) Form Time Refresh Missing / stale P
M1 Watching Which tokens am I tracking and what changed per card: symbol, chain, cap now, change since I started watching, lights lit, count of events since last looked, the newest event line with clock column of cards, ordered by newest change N T 1 min closed situations stay on the board, marked "closed ", until removed P0
M2 In How are the tokens I hold doing per card: H1 line (entry vs now, since entry, held for, from high), the top H2 event with clock column of cards, ordered by the most severe H2 event (H2 order) then recency N 30 s as H1 P0
M3 Out What happened after I got out per card: out time, cap at out, cap now and change since out, max cap since out column of cards, last 30 days T W 1 h n/a P2
M4 Alerts inbox and settings What reached me, and how list of alerts sent, newest first, with token and clock; settings: channel (Telegram bot / web push), quiet hours, per-rule switches for watching and in list + settings form N push "no channel set" P1
M5 Limits How much can I follow counts: watching n / 30, in n / 10 two numbers n/a on change n/a P1

Limits are proposed to keep alert volume sane and backend cost bounded; beta feedback sets the real numbers.

ASCII

+--------------------------------------------------+
| M5 watching 7/30 · in 2/10   M4 alerts (3) >     |
+----------------+----------------+----------------+
| M1 WATCHING    | M2 IN          | M3 OUT         |
| PONS $3.4M     | AI $420K       | GO out 09-24   |
| +18% since add | entry +17%     | $1.1M -> $0.4M |
| lights 5 · 4 new| held 1h52     | -64% since out |
| Pows bearish   | 23:40 Pows     |                |
|  23:40         |  turned bearish|                |
| ...            | ...            | ...            |
+----------------+----------------+----------------+

10. Data asks for T2 (infrastructure) and T4 (discovery)

Every field referenced above that is not clearly available in the audit's store table (DATA-AUDIT-2026-09-25 §2). "Have" says what the nearest existing store gives. Owner is who should plan it; T1 rules on rights for every external one.

ID Field Page / slot Have today Gap Owner P
DA-01 Tracked X accounts: mentions per token with clock and account Explore E5.5 (X light); room S5, S8 social-sweep has X files, dark since 09-18 (139 of 557 files Unauthorized) a working X source, a list of solid accounts, token resolution from text T4 P0
DA-02 Caller stance per token over time, and flips (bullish to bearish) room S8, H2 line 4, alerts discovery.jsonl has formatted calls only (callId, calledAt, callPrice, callMcap, snippet) read whole chats, classify each later message about a token as bullish / neutral / bearish / exit, keep the history T4 P0
DA-03 Expanded caller list Explore, S8, Record R2 5 channels, medians 1 to 7 tokens a day each more channels, with per-channel quality kept for R2 T4 P0
DA-04 Band-relative attention: mentions per hour for every open situation, so a rank within band can be computed S8 rank, Explore attention light, H2 line 7 events.sqlite has per-source events with token and clock a rollup: mentions per token per hour across sources, with the token's band at that hour T2 (rollup), T4 (sources) P0
DA-05 Early-buyer retention (first N buyers: holding / reduced / out) on BSC and Solana S7, H2 line 3, H1-adjacent holdings Robinhood only; 1,404 stub files "no tape" for BSC / Solana holdings on BSC (tape exists since 09-18) and Solana (tape restarted 09-25); until then tape-derived net positions labelled as such T2 P0
DA-06 Concentration clock: top-20 buyer share of supply, per minute from pool open S7 (profile A lead), S4 overlay telemetry/ per-minute per situation from the tape; no supply share field seen total supply per token, per-wallet net position from tape, share computed per minute T2 P0
DA-07 Tracked participant state per token (entry clock, size, holding / reduced / out, their median hold time and entry earliness) S9, H2 line 5 fomo-alerts events (182 tokens a day median) name trader and token; situations.jsonl carries actors position accounting per participant (the brand brief's "still maturing" item); behaviour stats per participant T2 (accounting), T4 (who is tracked) P1
DA-08 Holder count and top-10 / top-50 share S7 (profiles B and C), S10 not in any store provider (explorer API or indexer); cost per token per hour T2 with T1 P1
DA-09 Contract safety flags (mint / freeze authority, owner functions, honeypot test) S10, E5.7 not in any store a safety source (T1 names candidates) or own checks from bytecode / program accounts T2 with T1 P0
DA-10 Deployer identity and history (wallet, share held and sold, earlier tokens, earlier situations on this deployer) S10, S13, H2 line 2 tokens table has poolOpenAt, firstSeenAt; no deployer field deployer from the pool-create or token-create tx; a deployer index across tokens T2 P0
DA-11 LP events (add, remove, burn, lock) with amounts S10, S4 markers, H2 line 1 trades-bsc logs swaps; LP events not listed in the audit ingest mint / burn / transfer-to-dead / lock-contract logs on watched pools T2 P0
DA-12 Sniper share and bundle share in the first blocks S6a trades-bsc has block, wallet, size from pool open where the pool was watched from open watching from pool open (not from first sighting); funding-source link per wallet for bundles T2 P1
DA-13 Attention-to-money lag per call S8, Record R2 discovery.jsonl (calledAt) + trades-bsc (first buy after) on BSC a join, per call, on every taped chain; Solana and Base need tape T2 P1
DA-14 Narrative tags and narrative heat S11, S1, Explore narrative light none tagging (name / ticker heuristics, caller text, X text); heat rollup per tag T4 (tagging), T2 (rollup) P1
DA-15 Band-peer tables (band median volume / cap, peers by narrative) S3 profile C, S11 market-cache and ticks per watched token only a band universe per chain larger than our situations (provider market lists) T2 with T1 P1
DA-16 Volume 24 h and swap counts since open S3 ticks carry vol1h; tape can count swaps on taped chains 24 h volume field (provider) or tape sum; per chain T2 P0
DA-17 Market numbers for Base and other untaped chains Explore, S3 discovery row: "chain base not covered by the market broker (robinhood, solana, bsc)" broker coverage for Base at minimum (callers already call Base tokens) T2 P1
DA-18 Layer 1 aggregates: stablecoin supply by chain, ETF flows, chain DEX volume, TVL, bridge flows, BTC dominance, ETH/BTC, SOL/BTC, band volume / cap for large caps Brief B1-B8 none in stores external aggregates (T3 picks sources, T1 rights); one daily and one hourly job T2 with T3 P0
DA-19 Terminal volumes (Photon, Axiom, GMGN, BullX, Trojan, Padre) Brief B6 none router / fee-account measurement or public dashboards (T3) T2 with T3 P1
DA-20 Trenches daily rollups from our tape (volume, 2x count by band, top-5 share, new pools) Brief B8 lower bands, B9 trades-bsc, candles-onchain, candles-bsc; Solana tape restarted 09-25 daily rollup job; Solana tape kept running T2 P0
DA-21 Leverage aggregates (funding, OI, liquidations) Brief B10 none source choice T2 with T3 P2
DA-22 Light states per situation with decay and an event id per lit cell Explore E5.5, S5 situations.jsonl has reasons and actors; evidence empty on 47% of transitions a lights table: token, category, count, firstAt, lastAt, eventIds; no light without event ids T2 (store), T5 (rules) P0
DA-23 User state: follow states, entry, last-seen, alert rules, alert log board M1-M5, room S1, S2, H0-H3 none a small user store and accounts (T9 for beta access) T2 P0
DA-24 On-demand situation from search frame F2 none open a situation for any resolvable address on a covered chain T2 P1
DA-25 Alert delivery (Telegram bot, web push) with quiet hours and dedupe board M4, H3 Telegram bots exist for our own alerts a user-facing bot and per-user routing T2 P1
DA-26 Frozen room snapshots at close Record R3 case/ read model per token (28 live), strike snapshots snapshot every room at close (all slots), store forever (retention rule) T2 P1
DA-27 Outcome marks on honest fills per situation (1 h, 4 h, 24 h, 72 h, best within 24 h) Record R2-R4, S13 outcomes.jsonl (44k marks, no provider clock); ticks with 29% repeats marks keyed to situations with repeat and phantom classification applied T2 P0
DA-28 Source freshness feed for the frame and S14 F3, S14 events.sqlite sources and source_health tables; tape lag in trades-bsc receivedAt one small read model with lag per chain and per source T2 P0
DA-29 Zone states per strategy per token and per-band strategy record Z2-Z5, R6 deep-paper book and replay tool for one rule T6 defines strategies and forgiving win; a job that applies them per token T6, T2 P1
DA-30 Chart price source with display rights for untaped chains S4 on Solana (until tape), Base ticks and candles from DexScreener / CoinGecko / GeckoTerminal: licensed, derived-only per the 09-25 synthesis T1 to name a candle source we may draw; otherwise our own tape per chain T1, T2 P0

11. Open design questions for Thomas (ten, each with a recommendation)

# Question Recommendation
Q1 May the room draw a price chart for tokens not on our own tape, from a licensed source? Draw from our own tape wherever it runs (BSC, Robinhood, Solana once kept on) and pay for a candle source that allows display for the rest; never draw DexScreener raw. T1 decides which source.
Q2 Lights as eight fixed labelled cells, or only the lit ones? Eight fixed cells in fixed order, unlit shown empty and unseeable shown hatched. The empty cells are information ("no wallets yet"), and a fixed order is learnable in one visit.
Q3 Is "in" declared by the user or read from a linked wallet? Declared, one tap, optional entry cap, for the beta. Wallet linking adds a custody-adjacent trust question we do not need before people pay.
Q4 Rank the look list by lights, by recency, or by price move? Lights lit now, then lights lit in the last hour, then recency. No price-move sort; that is every screener.
Q5 How many rows on the look list? 25 by default, 50 ceiling, and the appearance threshold rises to 2 lights on busy days. Measured basis: ~400 tokens a day touch our sources, ~55 situations a day.
Q6 One skeleton with band profiles, or separate pages per band? One skeleton, three profiles, profile printed in the header and overridable. Separate pages would split the trader's habits and triple the build.
Q7 Zones: on by default, or a tab? A tab on the chart, off by default, with n and the win definition printed every time it is on. On by default reads as signals.
Q8 Is the Record public? Public summary (R1, R2 medians, R4, R7) and member detail (R3 rows, frozen rooms). Caller scorecards are the page people will share.
Q9 Where do alerts go? Telegram bot first (traders already live there), web push second. No email.
Q10 Does the hold view show money P&L? No. Change since entry cap and time held only. No size, no dollar P&L, no suggested exit. It keeps "trade management stays with the trader" literal.

12. What I did not do

Files

File What
T8-WIREFRAMES-V0.md this report
wireframes/index.html list of wireframes
wireframes/brief.html, explore.html, situation-first-hour.html, situation-day2.html, situation-30m-narrative.html, situation-hold.html, zones-overlay.html, record.html, board.html one per page and per room variant; 12-column grid on desktop, one column under 700 px
wireframes/shots/*-desktop.jpg, *-phone.jpg renders at 1440 and 390 px; all 18 checked for horizontal overflow (none)
scripts/T8-wireframes.py generates the HTML from one slot table; edit and rerun to change a layout
scripts/T8-shots.cjs renders the shots and checks overflow (NODE_PATH=<dir with playwright> node scripts/T8-shots.cjs)
scripts/T8-intake-volume.py, data/T8-intake-volume.txt the intake sizing behind the look-list row budget

Status: DONE