{"course":{"slug":"prompting","title":"Getting Real Work out of a Model","url":"https://addaly.com/learn/prompting","level":"beginner","track":"start"},"lesson":{"slug":"your-own-prompt-file","title":"The file that makes this compound","url":"https://addaly.com/learn/prompting/your-own-prompt-file","position":50,"of":72,"minutes":8,"answer":"Keep each working prompt with its date, its model, a \"use when\" line, the reason for its odd constraints.","keyIdea":"Keep each working prompt with its date, its model, a \"use when\" line, the reason for its odd constraints, and capitalised slots for the parts you swap: the notes are what make it usable three months later."},"body":{"format":"markdown","text":"## What \"prompt engineering\" means for most working people\n\nNot a job title. A small file of prompts that produced good output, with the specifics swapped out next time.\n\nEach one took you a few minutes to work out and saves those minutes every week afterwards. Nobody else can write them for you, because they encode what your work actually needs.\n\n## What goes in an entry\n\nNot just the prompt. Five things, and the extra four are what make it usable in three months:\n\n```\n## Supplier quote comparison\nLast checked: 2026-09-04 (Claude, worked)\nUse when: two or more quotes to compare on the same criteria.\n\nPROMPT:\nCompare the quotes below on: unit price, minimum order, lead time\nin days, delivery to [CITY]. One row per supplier. Quote the line\neach figure comes from. Where a quote does not state something,\nwrite \"not stated\" — do not estimate. Then one line: which is\ncheapest for [QUANTITY] units, and what would change the answer.\n<quoteA>[PASTE]</quoteA>\n<quoteB>[PASTE]</quoteB>\n\nNotes: without the \"do not estimate\" line it invents lead times.\nTested on: quotes from Adeyemi and Okonkwo, Aug 2026.\n```\n\nThe **date and model** tell you whether to re-test. The **use when** stops you reading five prompts to find the right one. The **notes** stop you removing the line that is doing the work: this is the entry people regret leaving out, because a bare prompt looks like it contains removable padding. The **slots in capitals** make it obvious what to replace.\n\n## Slots, not variables\n\n`[CITY]`, `[QUANTITY]`, `[PASTE]`. Capitals in brackets, because they are visible when you skim and impossible to leave in by accident: an unreplaced `[CITY]` in your output is obvious, whereas a leftover \"Nairobi\" from the last time you used it is not.\n\nThat is not a small point. The commonest error in reusing a prompt is a stale specific from its previous use, and slots are the only reliable defence.\n\n## Where to keep it\n\nA plain text file or a note. Not a system that needs maintaining.\n\nThe only real requirement is that you can find it in five seconds from wherever you work. A file you have to open a tool to reach is a file you will stop using, and a prompt library that goes unused is worse than none, because you think you have one.\n\nOnce you have more than about fifteen entries, split by area of work rather than alphabetising. You look for prompts by situation, never by name.\n\n## What to put in and what to leave out\n\n**In:** anything you have run three times. Anything that took more than ten minutes to get right. Anything with a non-obvious line in it: a constraint you discovered you needed.\n\n**Out:** one-off questions, anything you would rewrite from scratch anyway, and any prompt you cannot say what it is for. If the \"use when\" line is hard to write, the entry is not a tool yet.\n\n## Sharing it\n\nA prompt shared without its notes and its test items usually disappoints, which is the subject of a later lesson. If you hand one to a colleague, hand them the whole entry, including what it was tested on and what happens without the odd-looking line. The alternative is a colleague who tries it once, gets a mediocre result, and concludes the tools are overrated.\n\n## Why this is the durable part\n\nModels change and products change. What does not go stale is the accumulated knowledge of what your work actually requires: that lead times get invented unless forbidden, that your director wants the ask before the detail, that the export uses `totalAmount` and not `total_amount`.\n\nThat knowledge lives in your file, alongside the example set and the ten test items. Those three files are what this course is actually trying to leave you with. Everything else is explanation.\n\n## Start it with three\n\nDo not plan the library. Open a file this week and put three prompts in it: the one you used today that worked, the one you keep retyping, and the one that took you longest to get right.\n\nThree entries is enough to be worth opening, which is the only threshold that matters. A library nobody opens does not exist, and the way people end up with one they open is by starting far too small rather than by designing a structure first.\n\nAdd an entry whenever you catch yourself thinking *I worked this out before*. That thought is the signal, and it is reliable."},"check":{"question":"Why does an entry's \"notes\" line (recording why an unusual constraint is there) matter more than it appears?","options":[{"text":"It documents the prompt for compliance and audit purposes.","correct":false,"tempting":"A real benefit in some settings, but not the reason it changes outcomes for an individual user."},{"text":"It helps the model understand the intent behind the constraint.","correct":false,"tempting":"The notes are for the person, not the prompt; they are not sent to the model."},{"text":"Without it, an odd-looking line reads as removable padding, and removing it silently reintroduces the failure it was added to prevent.","correct":true},{"text":"It makes the entry easier to find when searching the file.","correct":false,"tempting":"That job belongs to the \"use when\" line."}]},"prev":{"slug":"the-regression-check","title":"The five items you keep forever","url":"https://addaly.com/learn/prompting/the-regression-check"},"next":{"slug":"prompts-that-break-later","title":"Why a prompt stops working","url":"https://addaly.com/learn/prompting/prompts-that-break-later"},"authorship":"Founder and series editor: Dr Mustafa Mun. Written by AI models to a written standard. Exam keys checked blind by a second model: 3,665 of 3,666 agreed. Academic review invited.","review":{"status":"Draft, open for academic review"},"cite":"Addaly (2026) 'The file that makes this compound', in Getting Real Work out of a Model. Edited by M. Mun. Addaly. Available at: https://addaly.com/learn/prompting/your-own-prompt-file (Accessed: 13 October 2026).","licence":{"name":"CC BY-NC-SA 4.0","url":"https://creativecommons.org/licenses/by-nc-sa/4.0/"},"reader":{"schema":1,"title":"The file that makes this compound","courseTitle":"Getting Real Work out of a Model","module":{"index":6,"slug":"from-lucky-to-reliable","title":"From lucky to reliable"},"answer":null,"keyIdea":"Keep each working prompt with its date, its model, a “use when” line, the reason for its odd constraints, and capitalised slots for the parts you swap: the notes are what make it usable three months later.","hero":{"kind":"plate","image":null,"alt":null,"caption":null},"blocks":[{"t":"h2","id":"what-prompt-engineering-means-for-most-working-people","runs":[{"text":"What “prompt engineering” means for most working people"}]},{"t":"p","runs":[{"text":"Not a job title. A small file of prompts that produced good output, with the specifics swapped out next time."}]},{"t":"p","runs":[{"text":"Each one took you a few minutes to work out and saves those minutes every week afterwards. Nobody else can write them for you, because they encode what your work actually needs."}]},{"t":"h2","id":"what-goes-in-an-entry","runs":[{"text":"What goes in an entry"}]},{"t":"p","runs":[{"text":"Not just the prompt. Five things, and the extra four are what make it usable in three months:"}]},{"t":"code","lang":"","label":"Text","text":"## Supplier quote comparison\nLast checked: 2026-09-04 (Claude, worked)\nUse when: two or more quotes to compare on the same criteria.\n\nPROMPT:\nCompare the quotes below on: unit price, minimum order, lead time\nin days, delivery to [CITY]. One row per supplier. Quote the line\neach figure comes from. Where a quote does not state something,\nwrite \"not stated\" — do not estimate. Then one line: which is\ncheapest for [QUANTITY] units, and what would change the answer.\n<quoteA>[PASTE]</quoteA>\n<quoteB>[PASTE]</quoteB>\n\nNotes: without the \"do not estimate\" line it invents lead times.\nTested on: quotes from Adeyemi and Okonkwo, Aug 2026."},{"t":"p","runs":[{"text":"The "},{"text":"date and model","b":true},{"text":" tell you whether to re-test. The "},{"text":"use when","b":true},{"text":" stops you reading five prompts to find the right one. The "},{"text":"notes","b":true},{"text":" stop you removing the line that is doing the work: this is the entry people regret leaving out, because a bare prompt looks like it contains removable padding. The "},{"text":"slots in capitals","b":true},{"text":" make it obvious what to replace."}]},{"t":"h2","id":"slots-not-variables","runs":[{"text":"Slots, not variables"}]},{"t":"p","runs":[{"text":"[CITY]","code":true},{"text":", "},{"text":"[QUANTITY]","code":true},{"text":", "},{"text":"[PASTE]","code":true},{"text":". Capitals in brackets, because they are visible when you skim and impossible to leave in by accident: an unreplaced "},{"text":"[CITY]","code":true},{"text":" in your output is obvious, whereas a leftover “Nairobi” from the last time you used it is not."}]},{"t":"p","runs":[{"text":"That is not a small point. The commonest error in reusing a prompt is a stale specific from its previous use, and slots are the only reliable defence."}]},{"t":"h2","id":"where-to-keep-it","runs":[{"text":"Where to keep it"}]},{"t":"p","runs":[{"text":"A plain text file or a note. Not a system that needs maintaining."}]},{"t":"p","runs":[{"text":"The only real requirement is that you can find it in five seconds from wherever you work. A file you have to open a tool to reach is a file you will stop using, and a prompt library that goes unused is worse than none, because you think you have one."}]},{"t":"p","runs":[{"text":"Once you have more than about fifteen entries, split by area of work rather than alphabetising. You look for prompts by situation, never by name."}]},{"t":"h2","id":"what-to-put-in-and-what-to-leave-out","runs":[{"text":"What to put in and what to leave out"}]},{"t":"p","runs":[{"text":"In:","b":true},{"text":" anything you have run three times. Anything that took more than ten minutes to get right. Anything with a non-obvious line in it: a constraint you discovered you needed."}]},{"t":"p","runs":[{"text":"Out:","b":true},{"text":" one-off questions, anything you would rewrite from scratch anyway, and any prompt you cannot say what it is for. If the “use when” line is hard to write, the entry is not a tool yet."}]},{"t":"h2","id":"sharing-it","runs":[{"text":"Sharing it"}]},{"t":"p","runs":[{"text":"A prompt shared without its notes and its test items usually disappoints, which is the subject of a later lesson. If you hand one to a colleague, hand them the whole entry, including what it was tested on and what happens without the odd-looking line. The alternative is a colleague who tries it once, gets a mediocre result, and concludes the tools are overrated."}]},{"t":"h2","id":"why-this-is-the-durable-part","runs":[{"text":"Why this is the durable part"}]},{"t":"p","runs":[{"text":"Models change and products change. What does not go stale is the accumulated knowledge of what your work actually requires: that lead times get invented unless forbidden, that your director wants the ask before the detail, that the export uses "},{"text":"totalAmount","code":true},{"text":" and not "},{"text":"total_amount","code":true},{"text":"."}]},{"t":"p","runs":[{"text":"That knowledge lives in your file, alongside the example set and the ten test items. Those three files are what this course is actually trying to leave you with. Everything else is explanation."}]},{"t":"h2","id":"start-it-with-three","runs":[{"text":"Start it with three"}]},{"t":"p","runs":[{"text":"Do not plan the library. Open a file this week and put three prompts in it: the one you used today that worked, the one you keep retyping, and the one that took you longest to get right."}]},{"t":"p","runs":[{"text":"Three entries is enough to be worth opening, which is the only threshold that matters. A library nobody opens does not exist, and the way people end up with one they open is by starting far too small rather than by designing a structure first."}]},{"t":"p","runs":[{"text":"Add an entry whenever you catch yourself thinking "},{"text":"I worked this out before","i":true},{"text":". That thought is the signal, and it is reliable."}]}],"check":{"question":[{"text":"Why does an entry’s “notes” line (recording why an unusual constraint is there) matter more than it appears?"}],"options":[{"text":[{"text":"It documents the prompt for compliance and audit purposes."}],"correct":false,"tempting":[{"text":"A real benefit in some settings, but not the reason it changes outcomes for an individual user."}]},{"text":[{"text":"It helps the model understand the intent behind the constraint."}],"correct":false,"tempting":[{"text":"The notes are for the person, not the prompt; they are not sent to the model."}]},{"text":[{"text":"Without it, an odd-looking line reads as removable padding, and removing it silently reintroduces the failure it was added to prevent."}],"correct":true,"tempting":null},{"text":[{"text":"It makes the entry easier to find when searching the file."}],"correct":false,"tempting":[{"text":"That job belongs to the “use when” line."}]}],"order":[0,2,1,3],"why":null},"prev":{"slug":"the-regression-check","title":"The five items you keep forever","minutes":8},"next":{"slug":"prompts-that-break-later","title":"Why a prompt stops working","minutes":8}}}