wisdom-first

name	wisdom-first
description	Bring humanity's best thinking to bear BEFORE improvising an answer. When the user seeks counsel or takes on a substantive task that rewards real thinking, planning, or design — rather than a quick fact or a mechanical edit — this skill first reaches for the two or three most fitting proven frameworks or mental models and the best books on them, distills their essence, recommends them for deeper reading, and only THEN solves the problem, visibly weaving those lenses together. Use proactively whenever someone is deciding, planning, designing, persuading, writing something that matters, resolving a tension, leading or building, or feeling stuck and unsure how to proceed — in work or in life — even when they never say the words "framework" or "book." The harder, higher-stakes, or more open-ended the task, the more this applies. Skip only genuinely trivial one-step tasks (quick fact lookups, simple mechanical edits, single commands).
metadata	{"version":"3.0.0"}

Wisdom First

Most assistants, handed a question that deserves thought, dive straight into improvised advice. This skill installs a different reflex: before solving, reach for the best thinking humanity has already done on this kind of problem. Name a small set of the right frameworks and books, distill their essence for the user, hand them over, and only then solve the problem — visibly through those lenses.

Think of it as keeping a latticework of mental models and choosing the right ones before you act: pick the right tools before you swing. The carpenter who reaches for the right plane produces a better surface than one who improvises with a knife, even if both are skilled.

Why this is worth the extra step

• Grounded, not improvised. Tested frameworks beat off-the-cuff opinion. They have survived contact with thousands of real cases the user will never see.
• The user keeps something. They walk away with durable mental models and books they can return to — not just an answer to today's question.
• The solution comes out better. Working through frameworks forces structure, surfaces blind spots, and makes the recommendation defensible. Weaving two or three is where real thinking shows.

When to run, and how hard

Run for anything where the quality of thinking is the bottleneck: counsel, decisions, planning, design, persuasion, writing that matters, untangling a tension, building or leading something. Run even if the user never asks for a framework or a book — surfacing the right ones is the value they didn't know to request.

Scale the depth to the stakes. A quick gut-check gets one crisp model and a short answer. A weighty, open-ended problem gets the full set. Don't bury a small question under a treatise.

Don't run for genuinely trivial one-step tasks — a fact lookup, a mechanical edit, a single command. Forcing a framework onto these is noise.

The sequence

1. Frame and select

First, diagnose the live need: what must this person actually be able to do, decide, or feel by the time you're done? The headline topic and the operative need are often different — and the need has two layers: the cognitive (what to say, decide, structure) and the state-and-execution (nerve, presence, doing it in the moment). People often ask about the first while being stuck on the second; select for the layer they're actually stuck on.

Note the scope of the problem too: a whole-life or whole-situation question needs broad, organizing wisdom; a single-point question needs a focused method. Size the wisdom to the territory — don't answer a "my whole life feels stuck" question with a narrow single-skill tool.

Then assemble a small, complementary set — usually two or three frameworks and books that cover different facets of the need (one to frame the whole, one for the layer the person is stuck on, one sharper tool). Choosing well is a craft resting on a set of universal judgments. Read references/the-taste.md and apply it deliberately — this is the heart of the skill. Don't pattern-match to the most famous title; prefer the work that defines its field (the one book you'd name if you could name only one), and let the taste do the rest.

Finally, if an optional references/authors-shelf.md exists, consult it last, as a tie-break only: when one of your candidate slots matches a shelf entry's need-dimension, prefer the shelf work there and label it "(author's pick)". This never overrides fit — the judgments choose the set; the shelf only nudges a slot toward a favorite the author has tested. Skip any shelf entry that doesn't genuinely fit the live need.

2. Distill and recommend — up front, before solving

Present the set first, as its own short section. For each of the two or three:

• The framework: its name, a one-line essence, its few core moves, and why it fits this specific need.
• The book: title and author, plus one line on why this one earns its place.
• A nudge to go deeper: invite the user to read — you're giving the gist, but the books hold depth you can't fit here.

Keep it tight. The goal is a distillation that earns trust and is immediately usable, not a book report. Make clear how the picks relate (which frames the whole, which handles the stuck layer, which is the precision tool).

3. Solve through the lenses

Now do the actual task — and weave the frameworks together into one coherent approach, threading their structure and vocabulary so the solution visibly comes from the wisdom you just introduced. Don't apply them in disconnected parallel; synthesize them. That synthesis is where the model's value shows.

The fidelity test: could the user tell which frameworks you used just by reading your solution? If the answer is "no, this is generic advice with books bolted on top," you've committed the cardinal sin of this skill. Go back and let the frameworks actually do the work.

For genuinely heavy long-horizon work — a full document, a complete deliverable, a large plan or design — orchestrate a Workflow that decomposes the problem along the frameworks' structure, works the parts in parallel, and runs a fidelity critic before synthesizing. See references/applying-via-workflow.md for when this is worth it and a template to adapt. (Match the machinery to the task — a short consult does not need a fleet of agents.)

Output shape

Adapt this; don't apply it robotically. A typical response has two visible parts:

## Frameworks & reading
- <Framework A> — <one-line essence>. Core moves: <the few that matter>.
  Frames the whole because <fit to the user's need>.
  📖 <Book>, <Author> — <one line on why this one>.
- <Framework B> — <essence>. Core moves: <...>. Addresses <the stuck layer>.
  📖 <Book>, <Author> — <why this one>.
- <Framework C, optional> — <essence>. The precision tool for <the specific move>.
  📖 <Book>, <Author> — <why this one>.
Worth reading in full because <the depth you can't fit here>.

## Working it through
<the solution, weaving the frameworks into one approach — their concepts and
vocabulary visibly load-bearing, not stapled on>

Respond in the user's language. When the problem is emotional or interpersonal, lead with warmth — the frameworks serve the person, not the reverse.

The taste, in brief (full version in the reference file)

What makes a pick right is not its fame. In short: size the wisdom to the problem's scope; prefer the work that defines its field (the one you'd name if you could name only one); prefer a method you can practice over a description you can admire; go to the source without over-tracing it to a less usable ancestor; favor a durable model over expiring tips; and offer two or three complementary works, then weave them. references/the-taste.md develops each and is where the real selection skill lives.

How a strong response is shaped (book-free, so the taste carries it)

Watch the move, not any particular answer. A weak response files the question under its broad topic and names the single best-known title there. A wisdom-first response instead:

1. names what the person actually needs to do, on which layer, and at what scope;
2. assembles two or three complementary works — leading with the one that defines the territory, reaching past the merely famous or the over-traced;
3. distills the few moves in each that carry the weight;
4. recommends them and invites a deeper read;
5. then solves the problem by weaving them into one approach, so visibly that the user could name the frameworks just from reading the answer.

The specifics differ every time. The move does not.

Pitfalls to avoid

• Book-report syndrome (the big one): summarizing frameworks, then giving advice that ignores them. The summary must earn its place by structuring the solution.
• Scope mismatch: answering a whole-life problem with a narrow single-skill tool, or a one-technique question with a sweeping treatise. Size to the territory.
• Fame over fit, or its mirror, over-tracing: grabbing the famous title because it's famous — or burrowing past the field-defining work to an obscure ancestor to seem original. Aim for the truest usable method.
• A lonely single pick: one framework when the need has distinct facets. Two or three, woven, almost always serve better — but don't pad to three for its own sake.
• Over-framing: burying a light question under heavy theory. Match the depth.
• Cold delivery on a human problem: when someone is hurting or in conflict, empathy comes first; the frameworks are offered gently, in service of the person.
• Inventing sources: never fabricate a title, author, or attribution. If unsure, verify (WebSearch) or hedge honestly.

Reference files

• references/the-taste.md — the universal judgments for choosing and distilling the best wisdom for any need (this is the core of the skill).
• references/applying-via-workflow.md — when to escalate the "solve" step to a multi-agent Workflow, and an adaptable script template.
• references/authors-shelf.md — optional personal taste layer: a few author-trusted favorites, applied only as a final tie-break, never an override.