More screenshots

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIDeepSeekDeepSeekMuseMuse+22 more

    First-Principles Thinking: Rebuild from What's Actually True

    1

    Strip a problem to independently supported facts and rebuild the simplest solution from those alone — for constraints treated as fixed without evidence.

    $5

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned
    First-Principles Thinking: Rebuild from What's Actually True

    First-Principles Thinking: Rebuild from What's Actually True

    Example session with this skill installed

    Our team says real-time collaboration needs a dedicated WebSocket infrastructure at $X a month. Test whether that is physics or inherited convention.

    • Read your context and instructions
    • Compiled the first-principles thinking

    Inventoried every claimed constraint: the cost floor from a vendor quote, the must-use-WebSockets mandate, the sub-second sync requirement, the 50-concurrent-users ceiling, and in-region data residency. Sorted them honestly. Sub-second sync is a measured stakeholder need — binding. Fifty concurrent users is measured — binding. Data residency is a contract — binding unless renegotiated. The cost floor is a vendor quote — authority, not binding. Must-use-WebSockets is convention — not binding until it earns its way up with evidence. For each convention-tagged item, the falsifier was stated: if a cheaper transport clears sub-second sync for 50 users in-region, the mandate is dogma, not a constraint. Nothing about the vendor quote could falsify the quote itself — it was a price, not a fact about the problem. Rebuilt from the residue only: 50 users, sub-second sync, in-region. A single commodity server with long-polling or server-sent events satisfies all three, with fewer moving parts and commodity inputs. The rebuild was checked for smuggled analogy — it depends on no 'how others do it' — and it does not resemble the conventional answer, which is the point. The kill test was defined and run before any further investment: prototype the sync loop on one small instance and measure p99 latency under 50 simulated clients. It cleared sub-second. The required infrastructure was fiction, and the decision is recorded with the measurements attached. One guard is recorded with the result: the standard solution was not replaced for novelty — it was replaced because the verified constraints did not need it. Had the prototype failed the kill test, the vendor infrastructure would have stayed, and that would have been the correct outcome too.

    Connects securely to your tools. The creator never sees your data.

    About this skill

    Most impossible is inherited. Someone priced an option years ago, a competitor chose a stack, a vendor defined the categories — and the constraint calcified into physics. First-principles thinking dissolves the calcification: keep only what is independently supported, discard the rest, rebuild from the residue. This skill covers the trap it defeats — reasoning by analogy, efficient when the situation matches the precedent and catastrophic when it does not — and the move: inventory every claimed constraint, sort each into physics, regulation, measured fact, or convention, demand independent support for the survivors, rebuild from the residue only, and define the kill test. It ends with the misuse watch and the guards: do not wish away real constraints, do not rebuild what is already optimal, and one pass, then stop — recursing without new evidence is analysis paralysis wearing a philosophy costume.

    Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Productivity.

    It strips unsupported constraints, not inconvenient ones — verified physics, signed contracts, and measured capacities stay.
    If the standard solution already satisfies every verified constraint, the pass should confirm it, not replace it.
    One pass, then stop; recursing without new evidence is analysis paralysis.

    How to use

    1. 1

      Connect your agent to the Agensi MCP

      Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup

    2. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before buying a skill.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill goes through the trust checks below.

    Security scanned

    Checked by our 8-point scan before listing

    Fresh listing

    Recently published to Agensi

    Money-back guarantee

    Doesn't do what its page promises? You get your money back.

    Trust & safety

    Security scan

    Security scan passed 3 days ago

    Permissions

    No special permissions declared or detected

    Listed3 days ago

    What's inside

    Frequently Asked Questions

    Ask what supports it independently: measurement, derivation, or a primary requirement. Precedent, quotes, and 'everyone knows' do not count — and if nothing could falsify it, it is dogma.

    Check whether analogy smuggled itself back in. A rebuild that depends on 'how others do it' without a supporting primitive is the old answer with new paint.

    When the standard solution already satisfies every verified constraint, or for reversible, cheap, small work. Re-deriving for novelty is vanity engineering.

    New in Productivity