Each step has one prompt that runs the whole step. That step prompt already carries out every method inside the step; you do not assemble it from parts. The methods are listed under each step so you can see what the step prompt does, and each one is also available as its own small prompt for the times you want to run just that move in a chat. To run the whole method across every task without pasting anything, use the system prompt.
Turn a topic into a question you can actually research, before any searching starts.
Work through the framing of this research with me before searching. Use the topic we have been discussing, or ask me for it if there is none yet. If a later step sent us back here because the evidence broke the original framing, do not start over: say what specifically changed, and reframe only what has to change. 1. Restate it as the decision it informs, the questions needed to make that decision, and the evidence that would answer them. 2. Tell me whether the question as posed is the right one, and whether a nearby question would be more worth researching. 3. Once we agree the question, mark which parts are background and which the conclusion will depend on, so we know where to go deep. Wait for me to confirm the framing before moving on. Once I confirm, end your reply with a short block I can carry into the next step: CARRY FORWARD Decision this serves: [one line] Settled question: [one line] Load-bearing (verify hard): [list] Background (report but do not rely on): [list]
Collect material and work out what can be trusted, one item at a time.
Research this question, working to these rules the whole way through. Start by restating what you are inheriting from the framing step: the decision, the settled question, and the load-bearing versus background split. Work from that. If it is not present in the chat, ask for it rather than inventing scope. If your research gives you reason to change any of it, say so explicitly instead of quietly re-scoping. - As you read, flag leads worth following: named people, referenced but unlinked documents, conflicting numbers, and things that should be present and are missing. - Treat weak sources (marketing, social posts, unattributed articles) as leads only. Take the checkable detail out and confirm it against a primary document, official body, peer-reviewed work, or named specialist reporting. Do not accept another version of the same weak source repeating the claim as confirmation. - Record every claim that matters in a table: the claim, whether the conclusion depends on it, its status, its best source and what kind it is, and what you searched. - Do not rest anything important on a single source when a second could be found. After three failed searches for a detail, mark it unconfirmed and move on. - If what you find undoes something the framing assumed, or makes the settled question the wrong one to be asking, stop and say so. Recommend going back to the framing step rather than researching on against a brief the evidence has already broken. Show me the table as you build it, and tell me what you could not confirm. End with a short process note: PROCESS NOTE Inherited: [the decision and load-bearing split you worked from] Did with it: [verified which load-bearing claims; what you left as background] Drift: [none, or what changed; if it breaks the framing, say to go back to framing] Handing on: [the confirmed claims and the still-open ones the next step needs]
Produce findings that are not sitting in any single source.
Help me analyse the material we have, rather than summarise it. I want findings that are not obvious from any single source. Start by restating what you are inheriting: the settled question and the confirmed material the gathering step handed on. Analyse only what was confirmed; do not revive claims that failed verification. If something you need is missing, ask rather than filling the gap with unverified material. - Where two things can be set against each other (a policy against a lifecycle, one option against another), lay them out and read the shape of the result: where things spread, bunch, crowd or leave gaps, and what that implies. - Where the two are peers, tell me where they differ: what each treats as central, where one is specific and the other vague. - If a comparison does not hold together, say so and suggest a better one rather than forcing it. If the comparisons keep failing to hold, treat that as a sign the question was framed wrongly and recommend returning to the framing step, rather than forcing findings out of material that will not support them. For each finding, tell me what in the material supports it. End with a short process note: PROCESS NOTE Inherited: [the question and the confirmed material you worked from] Did with it: [what you compared, what shape you read] Drift: [none, or what changed; if comparisons will not hold, say to go back to framing] Handing on: [the findings the conclusion should be built from]
Test the finding and put it in a form you can trust.
We are ready to conclude. Work from the research in this chat, and work through these in order. Start by restating the whole chain you are inheriting: the decision, the settled question, the confirmed claims, and the findings from analysis. Build the conclusion only from confirmed material and only in service of the stated decision. If the conclusion you reach does not actually serve that decision, say so rather than answering a different question. 1. State what the verified material actually supports. Not the direction most of the material seems to point, but what has actually been confirmed. 2. Try to disprove that conclusion. Search for evidence against it, worded to find the opposite. Address the strongest objection you find. 3. Write the conclusion so it names the specific claims it rests on, each confirmed. 4. Check the result: every fact traceable to a good source, anything unconfirmed marked as such, the contrary case addressed. 5. If a claim the conclusion needs fails its check, or the contrary evidence shifts a claim others were built on, do not conclude over it. Say which earlier step has to be reopened (the gathering of that claim, or the framing if the whole question is affected) and stop there. Show me the conclusion and the claims it depends on. End with a short process note: PROCESS NOTE Inherited: [the decision, question, and confirmed claims you built from] Did with it: [how the conclusion follows from them] Drift: [none, or where it departs; if a claim failed, say which step to reopen] Rests on: [the specific confirmed claims the conclusion depends on]
Turn the concluded research into the layered post people will actually read.
Turn the concluded research into a report for internal sharing. Work from the research in this chat, or ask me where it is. Start by restating what you are inheriting: the decision this served, the confirmed conclusion, and the claims it rests on. Report only what was confirmed. Anything unconfirmed is named as such or left out, not smoothed into the prose. Produce three layers, each doing only its own job. 1. Top post. The specific finding, stated as a fact, and one sentence on why it matters to the people who will read it. Name the actual development, not a category. Nothing else goes here. 2. Thread. The questions a reader has right after the finding: why now, what has actually changed, who should care, and any flag. Short paragraphs, the load-bearing word first. One comparison or piece of context if it earns its place. Do not walk through the source material in order. 3. Linked document. The full treatment, for the reader who wants depth: the finding stated first, then the evidence with its sourcing and how well each claim is stood up, the second-order implications, and the limits. Front-load the conclusion within each section rather than building up to it. Keep the finding the same across all three layers; the layers differ in depth, not in what they claim. State the conclusion first, never buried in a closing line. Plain declarative sentences. End with a short process note: PROCESS NOTE Inherited: [the decision, conclusion, and confirmed claims you reported from] Did with it: [how the finding is carried across the three layers] Drift: [none, or where the report departs from the conclusion and why] Rests on: [the confirmed claims the report presents as fact]