Back to Resources
Prompt Library12 min read
The Singapore Prompt Library: 50 Production-Tested Claude Prompts
OTG's living library of Claude prompts tested across Singapore law firms, SMEs, accounting practices, and consultancies. Filter by industry, copy, ship today.
How to use this library
Each prompt is a starting point. Copy it, adjust the bracketed [VARIABLES] for your context, run it, and customise based on what you get back.
Save your customised versions in a private team doc. Over a few months you'll have an internal prompt library worth more than any course. Share useful new patterns back via haojun@ontheground.agency — we'll add them to the next update.
This page is the brand-moat hub. Most other OTG resource articles link back here. If you're new to Claude, also read Singapore-Context Prompting and PDPA Prompting Checklist for the foundational rules.
Legal — 10 prompts
1. Contract clause extraction — see AI Contract Review in Singapore.
2. Comparative review (vs. firm standard) — also in the contract review article.
3. Case-law summarisation — see Claude for Legal Research.
4. IRAC memo draft — same source.
5. Counter-argument anticipation — same source.
6. Privilege flagging for e-discovery — see AI E-Discovery in Singapore.
7. PDPA redaction pre-pass — see PDPA-Safe Claude Prompts.
8. Plain-language explainer for client — same source.
9. Email to opposing counsel — Singapore commercial English, polite-firm scaling.
10. Engagement-letter clause for AI use — see PDPA-Safe Claude Prompts.
SME ops — 10 prompts
11. Invoice chaser — tone-scaling by days overdue. Full prompt.
12. Customer inquiry triage — sales/support/complaint/spam classifier. Full walkthrough.
13. Meeting notes → action items — JSON-structured extraction.
14. Spreadsheet → board report — Full prompt.
15. Quote → proposal generator — turn rough notes into polished one-pagers.
16. SOP from a recorded walkthrough — Loom transcript → step-by-step SOP.
17. JD from messy notes — Full prompt.
18. CV screen and score — same source.
19. Email reply summariser — long thread → status + next action.
20. Daily ops digest — overnight metrics → 200-word morning brief.
Audit and accounting — 10 prompts
21. Workpaper completeness review — Full prompts in Claude for Auditors.
22. Analytical procedures commentary — same source.
23. Sample summary writeup — same source.
24. Exception summary for partner — same source.
25. Management letter draft — same source.
26. Audit-committee report draft — same source.
27. Tax-computation review aid — see Claude for CFOs.
28. Cash runway memo — same source.
29. KPI dashboard narrative — same source.
30. Investor update letter — same source.
Consulting and advisory — 10 prompts
31. Pyramid principle structurer — Full prompt in AI Slide Generation.
32. Slide-by-slide spec writer — same source.
33. SCR storyline builder — same source.
34. Chart-picking assistant — same source.
35. Action-title trimmer — *"Trim each title to ≤14 words. If it can't, the underlying idea isn't sharp — restate."*
36. Strategic option scorer — score 3 strategic options on 5 criteria with weighted scoring.
37. Competitive benchmark generator — produce a clean comparison table from raw research.
38. Stakeholder map builder — given an org context, produce a stakeholder map with positions and influence.
39. Risk-rated recommendation — re-cast a draft conclusion as traffic-light risk-rated advice.
40. Plain-English board summary — for any analytical output, produce a 100-word non-specialist board summary.
Cross-cutting — 10 prompts
41. PDPA redaction pre-pass — Full prompt.
42. Singapore commercial English filter — *"Re-write the above in Singapore commercial English. No 'utilize'. No 'leverage' as a verb. No 'going forward'. Reuse 'we' or 'the firm' rather than 'our team'."*
43. Counter-argument generator — for any draft conclusion, generate the three strongest objections.
44. Plain-language reframer — for jargon-heavy text, produce a non-specialist version preserving accuracy.
45. Tone scaler — *"Re-write at three tone levels (warm, neutral, firm). Use the same content."*
46. Decision-quality check — given a decision and rationale, produce a structured pre-mortem.
47. Singlish-aware customer service — handle messages in Singlish or mixed-language inputs naturally.
48. Q&A FAQ generator — for any topic, generate the 8 most-asked questions with crisp answers.
49. Glossary builder — for jargon-dense documents, extract terms and define each in plain English.
50. The "explain like I'm a CEO" prompt — *"Explain [topic] in 5 sentences for a CEO who has no time and wants to know: what is it, why does it matter, what are we doing about it, what's the risk, what's the ask."*
Where to start
If you've never used Claude, start with Claude Code in Singapore: A Non-Developer's First-Week Guide. For Singapore-context rules, read Singapore-Context Prompting: PDPA, IRAS, MAS, EDG. For PDPA safety, read PDPA Prompting Checklist.
For tools to run these prompts on autopilot, see 7 Internal Tools Every SME Should Build with Claude Code.
For custom training and tooling for your team, book a free 30-minute call. OTG runs vibe coding workshops for SG law firms, accounting firms, and SME teams — partner-led or associate-led, half-day to two-day formats.
Frequently asked questions
Are these prompts free to use?
Yes. Copy, adapt, deploy. Attribution appreciated but not required. We update this library quarterly as patterns evolve.
Will the same prompt work in ChatGPT?
Most translate with minor adjustments. Claude responds better to XML tags ([Anthropic's recommendation](https://docs.anthropic.com/en/docs/build-with-claude/prompt-engineering/use-xml-tags)); ChatGPT prefers Markdown. Output format may vary.
How do you choose which prompts to include?
We include prompts that have been used in real Singapore engagements, refined across multiple iterations, and produce reliable output. We exclude prompts that worked once and never again.
Can I get a downloadable version?
Not yet — the library lives on this page (canonical, always-fresh). If a downloadable PDF would help your team, email haojun@ontheground.agency and we'll consider including it in a future update.
How is this different from Anthropic's prompt library?
[Anthropic's prompt library](https://docs.anthropic.com/en/prompt-library/library) is global and generic. Ours is Singapore-context — PDPA-aware, IRAS / MAS / EDG vocabulary, Singapore commercial English, real SG SME use cases.
PreviousClaude for CFOs: From Raw Ledger to Board PackNext PDPA Prompting Checklist: 8 Rules Before You Hit Send
Want to Apply This to Your Business?
We're a Singapore AI development and automation agency. Let's discuss how we can help solve your specific challenges.