Agent Basics — Meal Planning

Adapted from the DGF Agent Basics series — consumer-first guides for working with a personal AI agent. Pick only what you need.

> Pick only what you need. This skill is individual. This is one person's thoughts, offered as a pre-configured help, not a must. Take only the value you need from it, prune it to what fits you, and leave the rest. You do not have to use what the author thinks, and you do not have to take any other skill with it.

Summary: A meal plan for weight loss or muscle building, one lane at a time, with sleep, hydration, and micronutrients through food, tracked daily, and a perfect-10 score. Never recommends any supplement, ever. Optional.

Use this skill when the user wants a meal plan for weight loss or for muscle building. It follows the same rules as a workout plan: start as low as possible, or as high as possible in the loss case below, grow or shrink by one lane at a time, and track everything. Meals, sleep, and workout are the three biggest factors in any training, along with hydration and micronutrients. This skill is education and planning only. It is not medical advice and it is not nutrition advice for a specific medical condition.

This skill works on its own, with no other skill needed, and no knowledge of any other skill assumed.

Startup and how to use

Start by getting your own personal agent. Get your own personal agent and set it up. Bring your own accounts and logins. Use nothing that belongs to anyone else.

Tell your agent what you care about in plain words. Name your business, your work, your family, your hobbies, or any topic you want it to follow. Talking is the setup. You do not have to set up a system first.

Ask for the plan first, before any work runs. Say the task, then ask for the exact steps it will take, what it needs from you, and anything it cannot do. Wait for you to approve or correct the plan. Review the plan like a checklist and fix a wrong step or add a missing step before work starts.

Check that your agent understood you, before it runs. Ask it to show back what it thinks you mean and what output it will give you. Correct its repeat back right there when it does not match what you pictured.

Work in small single steps. Count the steps first. Give one step, check it, then give the next step. A ten step run is a huge run for an agent, so break it down and keep each step to exactly one thing.

Show it once, then save it by name. This is how you create a new task or a new skill. Take over the browser and do the task once yourself while your agent pauses. You drive alone and your agent does not watch your clicks. There is no record button to press. After you hand control back, tell your agent the steps in chat and tell it to save them by name, such as save this as Fun Things. Later, say pull up Fun Things and it picks up where you left off.

Correct it in small plain words when it does it wrong. Do not start over. Say what to change and say to remember that correction. That preference sticks and it gets better for you specifically over time.

Let it keep learning your interests every day. Your Feed looks like a newspaper with articles and topics related to what you told it you care about, updated daily. Your Ideas look like a lightbulb with suggestions for things you might like. It is not mind reading. It learns from what you actually talk about.

Use easy first and deep only when you want it. Never require a deep setup to use a basic task. Keep the wording in plain benefits and do not lead with virtual machine or VM. Say it has its own browser and workspace built in, you do not have to set anything up, it keeps your logins safe, and it saves and remembers your stuff.

Diverse, not concrete: let the user prune

Do not make this skill concrete, narrow, or one size fits all. Present a diverse set of options first, across price, style, risk, format, and approach, so different people can see themselves in it.

Let the user prune. Ask what to cut, what to keep, and what to change, and then narrow to what they kept. Pruning is the user's job, not your agent's guess. Do not prune for them before they have seen the options.

Do not lock a single choice early. Offer at least two or three meaningfully different options at each decision point, with the trade for each, unless the user already locked that choice themselves.

Keep options that do not fit one person, because they may fit another person using the same skill. A skill that only works for one kind of person is too concrete to reuse.

When the user prunes, save the pruned version by name when they want to repeat it, so their version starts from what they kept, not from the full diverse set again.

The three biggest factors, and safety first

Meals, sleep, and workout are the three biggest factors in any training, along with hydration and micronutrients. A meal plan that ignores sleep and hydration, or a workout plan that ignores meals, is incomplete in this method. Track all of them, together, every day.

This skill does not give medical or nutrition advice for a specific condition, and does not clear anyone to change their diet, their sodium, or their training. Ask the user, before any plan is built, whether they have a medical condition, a clinician's limit, an allergy, or a medication that affects food, sodium, micronutrients, or training. If they do, tell them to check with their clinician or dietitian first, and build the plan only inside the limits they state.

Protein, calorie, sodium, and micronutrient numbers in this skill are the author's method, stated by the author, not a universal prescription and not a guarantee. If a clinician gave the user a different limit, the clinician's limit wins, every time.

Setup: goal, weight, and plan length

Ask for goal first: weight loss, or muscle building, in plain words. The two lanes run in opposite directions, so a plan without a goal cannot be built.

Ask for total body weight, in pounds, and for plan length, in weeks. Use total weight, not muscle mass, because muscle mass is harder to figure, and the author's method uses total weight for its protein figures.

Ask for foods the user likes, foods they will not eat, allergies, diet style they want to follow, cooking time, and budget. Any diet style is okay in this method, as long as the rules below stay the same. Offer diverse styles, and let the user prune to theirs.

Ask for current meals, in plain words: what they eat, how much per meal by weight when known, and when they eat. A plan that does not start from what they actually eat will not be followed.

The rules, same as a workout plan

Meals must be very locked in, just as important as the workout. A locked in meal plan means meals are planned, weighed or measured, and logged, not guessed each day.

Start as low as possible, and grow over the set period of time, in the gain lane, and start as high as possible, and shrink by one lane at a time, in the loss lane, as below. The sequence matters more than any single day's numbers.

Increase or decrease by one lane at a time, never two at once. That is the same one factor per week rule as a workout plan, applied to food. When one lane is being changed, the other lanes hold.

Every week must move in some way, in the direction of the goal, unless illness, injury, or a clinician's limit says otherwise. Record a missed week plainly, and do not double the next week to catch up.

Lane 1: muscle gain

Start as low as possible. Start with the lowest amount of calories, protein, carbs, and fats that the user can possibly use in the beginning to gain weight, and that they can actually eat every day. Do not start at the maximum.

Add protein slowly, over time, as the first lane. Increase protein by a small, stated amount each week, and log it, until protein reaches its maximum requirement.

Maximum protein, in the author's method: 2 grams per pound of total body weight would be the absolute maximum ever needed, and the author states most people probably would not be able to handle that. 1 gram per pound of total body weight would probably be the maximum for most people. Example, in the author's words: a 200 pound or 250 pound person eating 250 grams of protein a day is a lot of food, and gaining weight on that amount is tough. Present these figures as the author's ceilings, not as targets every user must hit, and let the user prune to a protein amount they can actually eat, inside any clinician's limit.

Once protein hits its maximum, stop increasing protein. Then pick a primary driver: either fats or carbs, but never both at once. Increase the chosen primary driver, while protein holds, and the other lane holds. If the user later wants to switch primary drivers, switch one lane at a time, with the user's approval, and record the switch.

Follow the same sequence as a workout plan: start as low as possible, grow over the set period of time, one lane per week, until the plan's maximum is reached in the final week, never at the start.

Lane 2: weight loss

Start as high as possible. Calories need to be as high as they possibly can be at the start, while still moving toward loss, so the user has room to shrink later. Do not start at the lowest calories.

Take protein down to as low as possible first, in the author's method: around half a gram per pound of total body weight. That amount, in the author's words, should be enough to hold muscle, because most people are probably not training hard enough to lose muscle at that level. Present that figure as the author's floor, not as a guarantee, and keep it inside any clinician's limit.

Then take from fats and from carbs, one lane at a time. Pick one lane, and drive it down as low as it can possibly go, while the user still feels okay. Then pick the other lane, and drive that one down, while the first lane holds. Never drive both lanes down at once.

Feel good while doing it, in the author's words. If the user is having problems, or feeling hungry, keep the weight of food the same per day, and change what fills that weight to lower calorie foods. Example: if the user eats 22 ounces per meal, keep 22 ounces per meal, and fill it with lower calorie, heavier, fibrous foods.

Use volume swaps, in diverse options the user can prune: switch from rice to potatoes, add more vegetables, add fibrous foods, and add other lower calorie foods that still have weight. The weight per meal stays the same, the calories per meal go down, and hunger is easier to manage.

Take a lane down as low as it can go while feeling okay, not as low as it can go while feeling miserable. Feeling okay is a check, every week, not an afterthought. If feeling okay fails, hold that lane, and tell the user plainly, instead of pushing further.

Sleep, hydration, and micronutrients: maximise and track them

Maximise sleep every single day, maximise hydration every single day, and maximise micronutrients every single day, within any clinician's limit, whether losing or gaining weight. In the author's method, these help the most, alongside meals and workout.

Say micronutrients, not just sodium and minerals, because micronutrients give a better picture of what the user is actually doing. Micronutrients include everything from branched chain amino acids to minerals to vitamins. The more micronutrients ingested through food, the healthier the user is likely to be, and the more complete the diet is, because meeting requirements through food is itself very hard and very planned.

Maximise micronutrients means meet micronutrient requirements, through food. Meeting minimal requirements for all micronutrients is a very tough thing to do, and it has to be very planned and dedicated to actually consume enough food to meet them. That difficulty is the point of tracking them.

Micronutrients come through food, never through supplements, in this skill. Never recommend any supplement, ever, in this skill, in any message, or in any app built from this skill. That rule is absolute for the general, everyday people use this skill is built for.

Eating food is different from taking a supplement. Saying eat the most micronutrients possible through food is never going to hurt anyone, because meeting requirements through food alone is a very difficult thing to accomplish, and overdoing micronutrients through food is next to impossible. A person would basically just gain weight from eating way too much food, and even that would still be hard to actually accomplish.

Supplements are a slippery slope, and they are not something to tell someone about in general. A supplement need is a specific case, not a general case, and it needs a professional, specific instructions, blood work, and workups, in a lot more detail, with knowledge of what that actual person needs, not a vague guess. This skill is a general, everyday people skill, not a specific case skill.

If a user asks about a supplement, do not recommend one, do not name one, and do not guess whether they need one. Say plainly: this skill never recommends supplements, a supplement need is a specific case that needs a professional with blood work, and that decision is outside this general skill. Do not exceed a clinician's sodium limit in the name of maximising it, and do not use this skill to work around a clinician's limit.

Track sleep exactly: hours slept per day, and sleep quality in the user's own words, every day. Knowing how much sleep per day shows what variables can be changed.

Track hydration exactly: amount of water and fluids per day, in ounces or litres, every day. Knowing hydration per day shows what variables can be changed. Micronutrients aid hydration, in the author's method, because electrolytes and minerals are part of how the body holds and uses water, so track them together, not separately.

Track micronutrients with the best case tool the user will actually use: an app that allows micronutrient tracking, such as Cronometer, or another app that tracks micronutrients. If the user does not have such an app, offer Cronometer first, check whether they want to load it, and use whatever tracking they will actually do, rather than no tracking at all. It is hard for people to track all of these things, and that is exactly why your agent helping them matters: with help, they can put together a plan that actually meets micronutrient requirements, instead of guessing.

Track micronutrients in plain words, by category, when full tracking is too much at first: vitamins, minerals including sodium and electrolytes, and amino acids including branched chain amino acids. Start with the categories the user can track, and expand to full micronutrient tracking over time, one category at a time, with the user's approval.

Anything not tracked is something that could put the user behind. If a factor is not in the log, it cannot be changed on purpose, because there is nothing to change from.

Track everything, in a Google Sheet, and ask your agent back

Put everything in a Google Sheet, or in the spreadsheet of the user's choice, and track everything that can be tracked: meals with calories, protein, carbs, fats, and weight per meal, workout with its measure and amount, sleep hours, hydration amount, micronutrients by category, and a daily score.

Score each day out of 10, called a perfect 10 when every factor for that day was hit: meals locked in, workout done, sleep maximised, hydration maximised, micronutrients tracked, and tracking complete. A day that misses a factor is not a perfect 10, and the miss is recorded plainly, not hidden.

The user should be able to ask your agent, at any time: what did I do this day, and your agent can tell them from the sheet, day by day. The user should be able to ask: how did my week go, how many days did I hit a perfect 10, and how many days did I miss, and your agent can answer from the sheet, with numbers, not guesses.

Review the week together, in plain words: how many perfect 10 days, which factors were missed most, what variable to change next week, and which lane moves next week, one lane only. If the user hit everything perfectly throughout, in the author's words, they should have success. If they miss, they will not have success, in the author's words, and the miss, not the person, is what to fix next week.

Save the meal plan by name when the user wants to repeat or adjust it, such as save this as Gain Plan, so the next version starts from the same lanes, table, and sheet, not from scratch.

Present the result

Lead with goal, total weight, plan length, starting amounts, and primary driver or loss lane, in plain words, then the weekly lane table.

State what you could not verify. State when a starting amount was not measured, a protein figure was estimated, a micronutrient amount was not tracked, or a clinician's limit was not stated.

Stop after a denial. Do not re-pitch a rejected food, lane, or diet style in new words.

Available connectors

[Google Sheets](/opt/hatch/connectors/google_sheets/CONNECTOR.md) for creating, reading, and editing the tracking spreadsheet.

Use your agent's connector and skill search, if it has one, to find other connectors and related skills.

If nothing here fits, look on the web for a maintained API or MCP server and set it up with your agent's custom connector setup, if it has one, or build a custom connector.