semantic-algos

Compile natural language, a loose composition or pipeline sketch, or an existing program.md into a readable Haskell-esque Sem program that composes this repository's semantic skills. Use when a user asks to express, design, revise, expand, or simplify a semantic program without running it.

Compile and execute natural-language requests, loose semantic pipelines, or existing program.md files as squishy Sem programs. Use when a user invokes $sem-run, asks to run a composition of semantic-algos skills, or wants every semantic function application isolated in a fresh subagent with standalone Markdown artifacts and a linked final result.

Transform a topic, claim, question, or situation into an original joke by leading the audience into one plausible frame and using a concise punchline to reveal a second compatible frame. Use when the user asks for a joke, one-liner, punchline, comedic reframing, roast, or short humorous treatment of an idea.

Transform a question, tension, memory, or belief into an original poem or song lyric whose meaning emerges through voice, imagery, sound, and recurrence rather than explanation. Use when the user asks for a poem, song lyrics, a lyrical treatment of an idea, or a literary form that embodies an emotional contradiction without resolving it.

Solve a problem by systematically importing solutions from other domains: abstract the problem's deep structure, find 3–5 domains that share that structure, study how each solved it, and translate the mechanisms back — with an explicit disanalogy check. Use for 'how do other fields handle this?' and stuck problems that feel like they must have been solved somewhere before.

Structure a choice between concrete options: elicit the real criteria and their weights, score each option honestly, run a sensitivity check on the weights, and separate what the numbers say from what the gut says. Use when choosing between 2–6 named alternatives (jobs, tools, vendors, apartments, strategies).

Transform a question into a short parable that embodies it without stating it — a minimal world with one strange rule that makes the question's tension physical, inhabitants who live out the possible answers, and an ending that refuses to adjudicate. Use when a question would illuminate more as story than as analysis, in the tradition of LeGuin's "The Ones Who Walk Away from Omelas."

Take a surface question and forge the question actually worth asking — diagnose what's broken in the asked question (smuggled answers, wrong level, false binary, self-protection), transform it, and return a better question WITHOUT answering it. Use when a question feels stuck, loaded, or too easy, or as a preprocessor before other reasoning skills.

Extract every assumption hidden in a claim, plan, or belief, classify each by load-bearingness and testability, and identify which single assumption would collapse the whole structure if false. Use when a plan feels solid but untested, or before committing to a decision built on unexamined premises.

Reason rigorously about "what if X had been different" — isolate the minimal change, propagate consequences forward step by step, and separate what almost certainly follows from speculation. Use for historical what-ifs, post-mortems, decision regret, and attributing causes ("did X actually matter?").

Use dynamic-programming-style reasoning as a general thinking tool for any problem, decision, strategy, explanation, or analysis. Decompose the question into overlapping subproblems, memoize subanswers, then synthesize a solution from the memo table.

Explain any topic at five escalating levels — High School, College, PhD, Philosopher, Gigabrain — where each level responds directly to the one before it, correcting, deepening, or reframing it. Use when the user asks for an "explanation ladder", a "five-level explanation", "ELI5 to galaxy brain", or any cascading multi-level breakdown of a topic.

Use first-principles thinking as a tool for any question, design problem, strategy, decision, or explanation where the user wants to reason from fundamentals instead of analogy, convention, or inherited assumptions.

Root-cause analysis preset — drill from a surface symptom to deeper causes, assumptions, or principles by asking "why" ~5 times. A convenience preset of the more general n-whys skill (which takes an explicit depth n).

Use the Golden Circle as a thinking tool for any question, pitch, product, strategy, writing, or decision where the user needs to clarify Why, How, and What.

Solve a problem backwards by asking how to guarantee failure, then negating. Use for goals, plans, designs, habits, or strategies where the failure modes are easier to enumerate than the success path — "how do I succeed at X" becomes "how would I guarantee X fails, and how do I avoid that."

Use the ladder of abstraction as a thinking tool for any question, explanation, writing task, strategy, or concept where the user needs to move between concrete examples and abstract principles.

Generate a why-chain exactly n levels deep for any topic, question, observation, problem, or idea. Invoke like `n-whys n=8 q=<my question>`. The general form of five-whys — use this when you want an explicit depth.

Explain any topic at three escalating Nietzschean levels — Camel, Lion, Child — where each level responds directly to the one before it, moving from burden-bearing understanding, to critical negation and freedom, to creative affirmation. Use when the user asks for a "nietzche ladder", "Nietzsche ladder", "camel/lion/child explanation", or any three-stage breakdown in the spirit of Nietzsche's metamorphoses.