Built HashNest’s static export, with source, lockfile, README, DESIGN.md and validation records.

Build/typecheck passed; 12 calculation tests, 17 browser checks and 16 design checks passed. Deliverable: 996,281 bytes, below 8 MiB.

Limitations: finite supported-coin coverage, Bitcoin-only ASIC data, and WebKit/Safari unverified because required system libraries were unavailable.

Not delivered: artifacts/BUNDLE.md, artifacts/DATA.md, artifacts/DESIGN-GUIDANCE-LICENSES.txt, artifacts/FONT-LICENSE.txt, artifacts/design-probe.json, artifacts/interaction-results.json, artifacts/screenshots/desktop-1440.png, artifacts/screenshots/mobile-390.png and 1 more. Under artifacts/, only this step's named outputs leave the machine.

Area	Result
Requirements (24)	20 met, 2 missing, 0 faked, 2 after review
Broken features	1
First screen at 390px	Yes; Calculate ends at 621px
taste@1	16/16 passed
Copy findings	1
Share card	1 present

Saved 3 verified findings: 1 high and 2 medium in .imd-findings.json. Site files remain unchanged.

Explicit requirement	Status	Evidence
“HashNest Mining Profitability Calculator”	Met	Browser title and HashNest branding
“small self-contained web tool”	Met	Static export with bundled assets
“Enter your GPU or ASIC”	Met	Graphics card / ASIC miner controls
“hardware model (dropdown with hashrate/power presets)”	Met	Hardware dropdown and algorithm preset disclosure
“electricity cost ($/kWh)”	Met	#electricity
“optional pool fee”	Met	#poolFee; blank becomes zero
“real-time profit”	Met	Supported coins refresh from public feeds; four live requests succeeded
“across every mineable coin”	Missing	Coverage dialog explicitly limits supported coins
“updated every 10 minutes”	Met	src/useMarket.ts refresh interval
“home miners…zero-setup, auto-optimized decisions”	Met	Default presets, automatic net ranking and switch recommendations
“hosted”	After review	Publication follows review
“mobile-friendly”; “Clean, fast, works on a phone”	Met	Responsive controls; narrow layouts and probe pass
“live daily/weekly/monthly profit”	Met	Summary totals and period controls
“top 10 mineable coins”	Missing	GPU displays ten; ASIC comparison contains only Bitcoin
“ranked by net USD”	Met	Descending net-profit sort in src/calculator.ts
“Auto-refreshes difficulty and price via public APIs”	Met	src/market.ts and refresh timer
“‘switch coin’ alert threshold setter”	Met	Current-coin selector and daily USD threshold
“No accounts”	Met	Calculator opens directly
“no install”	Met	Browser entry page
“one URL”	After review	Hosting URL follows publication
“Static”	Met	dist/ export
“runs in the browser”	Met	React entry and browser calculations
“no wallet”	Met	No wallet interaction
“no backend”	Met	Direct browser requests to public providers

HashNest is a home-mining comparison tool whose forest-green profit panel anchors a quiet, pale workspace.

• High — fidelity: Limited coin coverage misses the promised comparison, especially for ASICs.
• Medium — bug: Calculate scrolls the mobile profit summary completely above the viewport, reproduced in Chromium and WebKit.
• Medium — taste: Promotional closing copy becomes a tall, 102px-wide column at 390px.

Validation passed: 12 calculation/data tests, 17 Chromium checks and 7 WebKit checks. WebKit used controlled API responses. Findings include exact source snippets, fixes and measured reproductions.

Fixed all three findings and rebuilt dist/.

• Coverage: 34/33/33 GPU coins and nine priced ASIC coins, with daily-data labels.
• Passed 17 tests, 21 interaction checks in each browser, and all 16 design checks.
• Submission: 748 KB, below 8 MiB.

Responses are recorded in .imd-responses.json; documentation and validation results are updated.

Area	Previous review	This review
Requirements	Full counts unavailable	20 met, 2 missing, 0 faked, 2 after review, 1 needs requester
Broken features	1	0
First screen at 390px	Not recorded	Product and Calculate visible
taste@1	Not recorded	16/16 passed
Copy findings	1	0
Share card	Not recorded	Present; 1/1

Two prior findings are fixed. One high coverage finding remains, recorded in .imd-findings.json.

Explicit requirement	Status	Evidence
“HashNest Mining Profitability Calculator”	Met	Header branding, page heading and document title
“small self-contained web tool”	Met	Static dist/ export with local runtime assets
“Enter your GPU or ASIC”	Met	Hardware tabs and four presets
“dropdown with hashrate/power presets”	Met	#hardware and algorithm presets
“electricity cost ($/kWh)”	Met	#electricity
“optional pool fee”	Met	#poolFee; blank applies zero
“real-time profit”	Missing for supplemental ASIC records	Eight networks always use explicitly delayed daily data
“across every mineable coin”	Missing	Available, priced compatible GPU networks remain excluded
“updated every 10 minutes”	Met	Refresh interval verified with virtual time
“top 10 mineable coins”	Met for available records	Ten-row limit; nine currently priced ASIC networks; ten-row controlled case passes
“daily/weekly/monthly profit”	Met	Summary totals and period controls
“ranked by net USD”	Met	Descending net-profit calculation and table
“Auto-refreshes difficulty and price via public APIs”	Met	market.ts and useMarket.ts; 13 live requests succeeded
“‘switch coin’ alert threshold setter”	Met	Alerts.tsx; threshold, opportunity and disable states verified
“home miners…zero-setup, auto-optimized decisions”	Met	Default presets, automatic ranking and switch alerts
“only calculator”	Needs requester	Evidence supporting the exclusivity claim
“hosted”	After review	Publishing follows review
“one URL”	After review	Deployment address follows publishing
“mobile-friendly…works on a phone”	Met	Reflow and interaction checks in Chromium and WebKit
“Clean, fast”	Met	Immediate calculator surface, local assets, 16/16 design checks
“No accounts”	Met	Calculator opens directly
“no install”	Met	Browser-loaded static export
“Static, runs in the browser”	Met	Client-side React application
“no wallet”	Met	No wallet connection required
“no backend”	Met	Browser requests public providers directly

HashNest is a home-mining calculator whose dark-green profit summary makes the estimated return the visual focus.

• High — fidelity: GPU coverage remains incomplete. At 390×844, select RTX 4070, $0.10/kWh and 1% fee, calculate, then search Nirmata, Frencoin, Satoxcoin or Etica: each returns zero rows. BackPoW supplies usable priced records for all four, with algorithms already supported by the presets. Nirmata calculates to approximately $0.45/day. Extend supplementary normalization to compatible GPU records while preserving delayed labels. The limiting branch is market.ts:234.
• Fixed — bug: Calculate skipped the summary. Submission now focuses the results section; totals and loss warnings remain visible in both engines.
• Fixed — taste: cramped promotional tail. The strip is removed, the functional subtitle is present, and the phone footer gap measures 24px.

Validation passed: 17 calculation tests, 21 interaction checks in each browser, and all 16 taste checks. WebKit interaction checks used controlled API responses.

I reproduced the finding and fixed it. GPU rankings now include the priced BackPoW networks that the primary feed is missing, including the four coins the reviewer named.

What I confirmed first: with RTX 4070, $0.10/kWh and a 1% fee, the page showed "Show all 34", and searching Nirmata, Frencoin, Satoxcoin or Etica returned nothing. BackPoW's catalog lists 69 coins for the three GPU presets, and 31 priced, usable ones were absent from the comparison.

What changed:

• New coin list: src/market.ts has a second list next to the SHA-256 one, covering 29 GPU networks. Each record must report the algorithm it is listed under, or it is rejected.
• Left out on purpose: Kerrigan EquihashZero and PepePow report zero difficulty, and Tari merged mining only pays alongside Monero. Five coins with no price are also excluded. The reasons are in artifacts/DATA.md.
• No duplicates: if the primary feed already has a network, its five-minute data is used and the daily copy is dropped.
• Benchmarks: algorithms that already had presets (KawPow, ProgPowZ, Etchash, FiroPow) use them. The 18 that had none use BackPoW's figure for that exact card model. A card with no published figure for an algorithm gets no row.
• Units: the new algorithms show Sol/s, graphs/s or solutions/s instead of a misleading H/s.
• Labels and alerts: every added row is marked "Daily · Delayed" and can't trigger a switch alert. If one of these requests fails, that coin keeps its saved figures, and the failure doesn't mark the S21 view as saved.
• Bundled snapshot: 29 real records captured today were appended; the existing lines are unchanged.

Results:

• Coverage: RTX 4070 / 3070 / 3060 Ti now show 51 / 60 / 51 coins, up from 34 / 33 / 33.
• Nirmata: $0.8643 revenue and $0.4525 net per day on the RTX 4070, matching the reviewer's figure. It ranks 5th in the bundled data, with Satoxcoin 3rd.
• Live check at 390×844 against the real APIs: all 42 requests returned HTTP 200 with no console errors. The page showed "Show all 51", and each of the four named coins gave one "Daily · Delayed" row (Nirmata 7th with live data).
• Checks: typecheck and build passed, unit tests 19/19 (two new), browser tests 21/21 in Chromium, and the design probe 16/16. The 390 first screen is unchanged.

Limitations:

• WebKit wasn't rerun this round: its system libraries are missing on this machine and I didn't install system packages.
• Request count: each 10-minute refresh now makes 42 requests instead of 13. They are small CDN files, and the extra data only updates daily.
• Gaelium: BackPoW gives no ticker for it, so its name is shown in place of a symbol.

The export was rebuilt from the final source, and the docs have the new counts and results: README, DESIGN.md, artifacts/DATA.md, artifacts/validation.md and artifacts/BUNDLE.md. The answer to the finding is in .imd-responses.json, and nothing is committed.

Area	Previous supplied review	This review
Requirements	1 coverage gap; full tally unavailable	19 met, 0 missing, 0 faked, 2 after review, 2 needs requester
Broken features	Not supplied	0 reproduced
First screen at 390px	Not supplied	Pass; action ends at 642px
taste@1	Not supplied	16/16 passed
Copy findings	Not supplied	0
Share card	Not supplied	Present; metadata and image verified
Open findings	1 high	0

The previous GPU coverage finding is fixed. At 390×844, with RTX 4070, $0.10/kWh and 1% fees, Chromium and WebKit each completed all 42 API requests successfully. Nirmata, Frencoin, Satoxcoin and Etica each returned one matching row, marked “Daily · Delayed.”

GPU coverage is now 51 / 60 / 51 rows. Nirmata calculates to $0.45249718 net/day, matching the prior reproduction, and ranked seventh with the fetched primary data. Delayed records remain excluded from switch alerts.

Validation passed: 19/19 unit tests, 21/21 browser checks in each engine, and typechecking. No new reproducible defects were found.

Explicit requirement	Status	Evidence
“HashNest Mining Profitability Calculator”	Met	Page title, brand and calculator heading
“small self-contained web tool”	Met	Static export with local script, styles, font and images
“GPU or ASIC”; hardware dropdown with hashrate/power presets	Met	#hardware, hardware tabs and algorithm presets
“electricity cost ($/kWh)”	Met	#electricity; changing it recalculates power expense
“optional pool fee”	Met	#poolFee; blank input applies zero
“real-time” / “live” profit	Needs the requester	Uniformly live data requires a faster supplemental source; daily records are explicitly labeled
“across every mineable coin”	Needs the requester	Exhaustive coverage needs a complete inventory and usable data beyond supported catalogs; Data & coverage states the boundary
“updated every 10 minutes”	Met	Refresh timer verified with a virtual clock
“home miners” and “zero-setup”	Met	One-device presets; usable immediately without keys
“auto-optimized decisions”	Met	Highest-net recommendation, sorted rankings and switch opportunities
“hosted”	After review	Hosting is the publication step
“mobile-friendly”; “works on a phone”	Met	Chromium/WebKit interactions and reflow checks at 320/390px
“daily” profit	Met	Daily summary and table
“weekly” profit	Met	Seven-day summary and table
“monthly” profit	Met	Explicit 30-day summary and table
“top 10 … ranked by net USD”	Met	Ten GPU rows in descending net order; ASIC shows its nine currently priced networks
“Auto-refreshes difficulty and price via public APIs”	Met	Successful browser requests and interval checks
“‘switch coin’ alert threshold setter”	Met	Current-coin selector, USD/day threshold, save/disable and opportunity states
“No accounts”	Met	No account flow or authentication requirement
“no install”	Met	Entire interaction works in the browser
“one URL”	After review	Static entry page ready for publication
“Static, runs in the browser, no wallet, no backend”	Met	Browser calculations, local storage and direct public API requests
“Clean, fast”	Met	Immediate local calculation; 16/16 design checks; no missing local resources or script errors

The two data dependencies are disclosed limitations, not fabricated functionality; BackPoW documents daily updates.

The visual idea is a compact mining dashboard whose dark-green profit summary and lime primary action make the decision easy to find.

.imd-findings.json contains {"findings":[]}. Website files were unchanged.