Editor-selected existing cases and prompts, separate from AI news. Selection dates use Hong Kong time.
Five editorial adaptations check award fares, draft permissions, receipt refunds, image-ID joins and cropped-screen coordinates using fictional data. Not newly discovered real-world cases. Gemini web prompts are untested; cards include human checks, sources and regional limits.
Use Gemini web text chat to turn comparable quotes into a value-per-point table; flag missing FX or change terms rather than forcing a cheapest option. Prompts are untested; results require human review.
Prompt
Check [quotes] against [trip conditions]. Only with matching currencies and complete cash fees calculate (cash fare minus award cash fees)/points. Show formula, units, noncomparable terms and missing data. Use only [FX rate and date]; do not convert if absent or book anything.
Fictional fixture, not a real usage record: Same route, date, cabin and baggage: cash HKD 2400; A costs 20000 points plus HKD 800; B costs 18000 points plus USD 120, with no FX rate supplied. A is nonrefundable; cash is refundable but its fee is unknown.
Evidence limits and cautions
The original source provides inspiration, not validation of this exercise. Its toolchain, regional services and plans have not been reproduced. This card uses only the documented Gemini web text environment. The source repository describes cash/award comparisons and points-value rules. This card’s fictional quotes, missing-FX gate and human arithmetic are editorial additions. Original publication date: unknown; checked: 2026-11-12.
Original editorial adaptation, not a newly discovered use case or the source author’s verbatim workflow. Source-reported outcomes do not guarantee this exercise. Prompts were not tested. Adapted from this existing case: https://hkfrees.com/ai/case/reported-award-offer-comparability/?lang=en Use Gemini web text chat to turn comparable quotes into a value-per-point table; flag missing FX or change terms rather than forcing a cheapest option. Open this card’s title for the official environment-source links and human acceptance steps on its full case page.
- Platform and region: ordinary Gemini web text chat. Official lists include Hong Kong, Macau, Taiwan, Chinese variants and English. Documentation checked 2026-11-12; local connections were not tested.
- Plan and account: prefer an adult personal account. Standard free access is a starting point, with variable usage caps and no required subscription. Taiwan advertises NT$0/month free access; Hong Kong/Macau paid prices are unverified. No particular model or prompt count is guaranteed. Work, school and child accounts need separate checks.
- No original toolchain is connected. Web availability does not establish mobile, API, CLI or original-automation access. Never supply confidential data. Activity-off chats can still remain for 72 hours and receive safety review.
- Fictional quotes only, not live fares, a general points valuation or buying advice; no availability guarantee.
Original source: Travel hacking toolkit
Use Gemini web to map draft sentences to approved notes and flag changes after permission is withdrawn. Anonymization is not automatic permission. Prompts are untested; results require human review.
Prompt
Using [current permission list], map each clause in [draft] as [sentence ID → source ID → allowed/denied/unknown]. Keep only currently explicitly public facts. Remove withdrawn and internal facts; do not infer permission. Return a revision and reasons; publish nothing.
Fictional fixture, not a real usage record: N1 public: prototype completed. N2 internal only: Client A will trial it next week. N3 formerly public, now withdrawn: three feedback submissions received. Draft S1: Prototype completed and three feedback submissions received. S2: A client will trial it next week.
Evidence limits and cautions
The original source provides inspiration, not validation of this exercise. Its toolchain, regional services and plans have not been reproduced. This card uses only the documented Gemini web text environment. The source discusses separating brainstorming and private-agent notes to keep private logistics out of public drafts. Sentence IDs and the withdrawal fixture are editorial additions, not source-validated functionality. Original publication date: 2026-07-15; checked: 2026-11-12.
Original editorial adaptation, not a newly discovered use case or the source author’s verbatim workflow. Source-reported outcomes do not guarantee this exercise. Prompts were not tested. Adapted from this existing case: https://hkfrees.com/ai/case/reported-private-public-draft-boundary/?lang=en Use Gemini web to map draft sentences to approved notes and flag changes after permission is withdrawn. Anonymization is not automatic permission. Open this card’s title for the official environment-source links and human acceptance steps on its full case page.
- Platform and region: ordinary Gemini web text chat. Official lists include Hong Kong, Macau, Taiwan, Chinese variants and English. Documentation checked 2026-11-12; local connections were not tested.
- Plan and account: prefer an adult personal account. Standard free access is a starting point, with variable usage caps and no required subscription. Taiwan advertises NT$0/month free access; Hong Kong/Macau paid prices are unverified. No particular model or prompt count is guaranteed. Work, school and child accounts need separate checks.
- No original toolchain is connected. Web availability does not establish mobile, API, CLI or original-automation access. Never supply confidential data. Activity-off chats can still remain for 72 hours and receive safety review.
- Use synthetic notes only. Do not send genuine private data to an external AI merely to run a privacy check.
Original source: Cursor Customize for Brainstorming and a Personal Agent
Use Gemini web to reconcile fictional receipts, duplicate scans and refunds with source IDs, producing net totals and unresolved records, not tax deductions. Prompts are untested; results require human review.
Prompt
Structure [de-identified transaction text] into [source ID, transaction ID, currency, amount, type, linked ID, evidence, status]. Do not count duplicate scans twice. Offset a specific invoice only with linked refunds; list unmatched refunds separately. Distinguish matched-invoice net from all-transaction net outflow. Infer no tax eligibility; never sum different currencies.
Fictional fixture, not a real usage record: R1 invoice A HKD 100; R2 second scan of invoice A, same date/vendor/amount; R3 refund HKD 30 explicitly linked to A; R4 invoice B HKD 50; R5 refund HKD 10 with no matched invoice.
Evidence limits and cautions
The original source provides inspiration, not validation of this exercise. Its toolchain, regional services and plans have not been reproduced. This card uses only the documented Gemini web text environment. The source describes organizing transactions/receipts and spreadsheets with human review in a US tax context. Duplicate scans and refunds are editorial fixtures; source tax rules are not applied here. Original publication date: unknown; checked: 2026-11-12.
Original editorial adaptation, not a newly discovered use case or the source author’s verbatim workflow. Source-reported outcomes do not guarantee this exercise. Prompts were not tested. Adapted from this existing case: https://hkfrees.com/ai/case/reported-tax-records/?lang=en Use Gemini web to reconcile fictional receipts, duplicate scans and refunds with source IDs, producing net totals and unresolved records, not tax deductions. Open this card’s title for the official environment-source links and human acceptance steps on its full case page.
- Platform and region: ordinary Gemini web text chat. Official lists include Hong Kong, Macau, Taiwan, Chinese variants and English. Documentation checked 2026-11-12; local connections were not tested.
- Plan and account: prefer an adult personal account. Standard free access is a starting point, with variable usage caps and no required subscription. Taiwan advertises NT$0/month free access; Hong Kong/Macau paid prices are unverified. No particular model or prompt count is guaranteed. Work, school and child accounts need separate checks.
- No original toolchain is connected. Web availability does not establish mobile, API, CLI or original-automation access. Never supply confidential data. Activity-off chats can still remain for 72 hours and receive safety review.
- The source tax context does not establish HK/Macau/Taiwan tax rules. This is record organization; consult qualified help for actual tax decisions. Upload no account or identity identifiers.
Original source: Chris Blattman: Using Claude Code for Tax Season
Use Gemini web text tables to catch image mismatches, missing images and duplicate IDs after cards are sorted. No images are generated or uploaded. Prompts are untested; results require human review.
Prompt
Join [content table] to [image map] using [stable primary key], then apply [layout order]. Never join by row position. Flag missing values and conflicting duplicate keys; do not select the first duplicate. Preserve file IDs; do not claim visual inspection. Check [title-length limit] only if supplied.
Fictional fixture, not a real usage record: Content: P1 Mountain, P2 Sea, P3 City. Image map: P1→mountain.jpg, P2→sea.jpg, P2→lake.jpg; P3 absent. Layout order: P3, P1, P2.
Evidence limits and cautions
The original source provides inspiration, not validation of this exercise. Its toolchain, regional services and plans have not been reproduced. This card uses only the documented Gemini web text environment. The official xAI design guide describes Figma MCP asset placement and component consistency. The stable-ID/duplicate-key text exercise is an editorial addition, not a Grok or Figma test. Original publication date: 2026-08-24; checked: 2026-11-12.
Original editorial adaptation, not a newly discovered use case or the source author’s verbatim workflow. Source-reported outcomes do not guarantee this exercise. Prompts were not tested. Adapted from this existing case: https://hkfrees.com/ai/case/reported-template-card-data-binding/?lang=en Use Gemini web text tables to catch image mismatches, missing images and duplicate IDs after cards are sorted. No images are generated or uploaded. Open this card’s title for the official environment-source links and human acceptance steps on its full case page.
- Platform and region: ordinary Gemini web text chat. Official lists include Hong Kong, Macau, Taiwan, Chinese variants and English. Documentation checked 2026-11-12; local connections were not tested.
- Plan and account: prefer an adult personal account. Standard free access is a starting point, with variable usage caps and no required subscription. Taiwan advertises NT$0/month free access; Hong Kong/Macau paid prices are unverified. No particular model or prompt count is guaranteed. Work, school and child accounts need separate checks.
- No original toolchain is connected. Web availability does not establish mobile, API, CLI or original-automation access. Never supply confidential data. Activity-off chats can still remain for 72 hours and receive safety review.
- Filenames are fictional identifiers, not evidence of image rights. Check copyright and actual visuals before production.
Original source: Designing Grok Bot with Grok Bot
Use Gemini web to check screenshot-to-test-viewport coordinates with formulas and bounds. This is a paper exercise, not control of a real website. Prompts are untested; results require human review.
Prompt
Given [full-image size], [CSS viewport size], [crop origin and size] and [crop-local point], validate bounds, add crop offsets, then scale each axis. Show units, assumptions and result. Stop guessing if offset, coordinate frame or scale is missing. Execute no clicks.
Fictional fixture, not a real usage record: Full screenshot 1600×1000 pixels maps to an 800×500 CSS viewport. Crop starts at full-image (200,100), size 800×600. Target center is crop-local (400,300). Assume no scrolling, browser chrome or additional distortion.
Also test the same fixture with crop-local point (900,300); report whether it is out of bounds and whether calculation should proceed.
Evidence limits and cautions
The original source provides inspiration, not validation of this exercise. Its toolchain, regional services and plans have not been reproduced. This card uses only the documented Gemini web text environment. The source author explains Java browser control and scaling normalized model coordinates to a viewport. Crop offsets and image-pixel/CSS examples are editorial additions, not the source’s normalized 1000-grid. The visible publication date is November 2 despite the November 3 URL. Original publication date: 2025-11-02; checked: 2026-11-12.
Original editorial adaptation, not a newly discovered use case or the source author’s verbatim workflow. Source-reported outcomes do not guarantee this exercise. Prompts were not tested. Adapted from this existing case: https://hkfrees.com/ai/case/reported-browser-coordinate-calibration/?lang=en Use Gemini web to check screenshot-to-test-viewport coordinates with formulas and bounds. This is a paper exercise, not control of a real website. Open this card’s title for the official environment-source links and human acceptance steps on its full case page.
- Platform and region: ordinary Gemini web text chat. Official lists include Hong Kong, Macau, Taiwan, Chinese variants and English. Documentation checked 2026-11-12; local connections were not tested.
- Plan and account: prefer an adult personal account. Standard free access is a starting point, with variable usage caps and no required subscription. Taiwan advertises NT$0/month free access; Hong Kong/Macau paid prices are unverified. No particular model or prompt count is guaranteed. Work, school and child accounts need separate checks.
- No original toolchain is connected. Web availability does not establish mobile, API, CLI or original-automation access. Never supply confidential data. Activity-off chats can still remain for 72 hours and receive safety review.
- The formula holds only under the stated assumptions. It proves neither browser-automation availability nor safety. Do not use it for payments, deletions or unauthorized clicks.
- The original uses normalized model coordinates; this card instead uses explicit screenshot pixels and CSS coordinates. Do not mix these coordinate frames.
Original source: Driving a web browser with Gemini’s Computer Use model in Java