Convert a single platitude into concrete, testable actions, boundary conditions, and failure modes — transforming vague wisdom into executable guidance.
Skills in this repository
benjam3n/reasoningtool - Page 14
SkillsMP has collected 655 skills from benjam3n/reasoningtool. Open a skill to review its source and details.
benjam3n/reasoningtoolShowing 40 of 655 collected skills.
Analyze multiple platitudes as a system, detect contradictions, reconcile conflicts by identifying context conditions, and convert the set into a coherent, actionable policy.
Procedure for evaluating and improving plans. Checks if steps lead to goal, identifies missing or unnecessary steps, assesses sequencing and resources, and suggests improvements.
Managing procurement of components and materials — sourcing, ordering, tracking, receiving, and verification.
A 24-week rotating list of questions for weekly review.
Systematic procedure for improving what you notice. Slows down observation, separates what's present from what's interpreted, and surfaces what you normally filter out.
Elicit user preferences by presenting concrete trade-offs rather than asking open-ended questions. System does the work by generating options and letting user select.
Systematically prioritizes a list of anything using weighted criteria and appropriate frameworks (Eisenhower, MoSCoW, impact-effort).
Reason under uncertainty using Bayesian thinking, base rates, confidence calibration, and expected value. Avoid base rate neglect, conjunction fallacy, overconfidence, and other probability traps.
Review and improve the procedure registry schema using explicit definitions, evidence alignment, and non-regression.
Procedure for analyzing products or offerings. Identifies the user and their need, evaluates fit, compares against alternatives, and finds improvement opportunities.
Validate that a need is real before investing by finding behavioral evidence, quantifying demand, checking existing solutions, and assessing trajectory.
Analyze from the perspective of expert users. Identifies power user needs around efficiency, customization, and depth, finds friction points, spots missing advanced features, and assesses the learning curve.
Validate that a procedure is complete, executable, and all dependencies are satisfiable. Catches issues before execution.
Handler for goals focused on PREVENTING bad outcomes rather than achieving good ones.
Bridge thinking to real-world action — account for physical logistics, materials, locations, time, and failure modes that pure reasoning ignores.
Generate high-quality questions about identified guesses. Applies quality criteria and "why continue?" gate to prevent low-value questioning.
When you don't know something, the right question is worth more than hours of speculation.
Procedure for measuring outcomes that resist quantification.
Find the satisfying unresolved question that opens a document and order all questions in a dependency chain. Upstream of /w.
Extract structured QuestionRoute data (questions, routes, chains, regresses) from unstructured prose, skills, or markdown files. Prototype for structured skill output.
Orderings for managing queues — FIFO, priority queues, shortest job first, fair queuing.
Adaptive root cause analysis. Ask "why" until you hit the root. Sometimes that's 2 levels, sometimes 12. Stop when you find the cause, not at a fixed number.
Systematically trace symptoms back to their underlying root causes using structured diagnostic techniques. Supports context-adaptive variants.
Adversarial analysis — attack your own plan, argument, or system. Find the failure modes, exploit the weaknesses, identify what an intelligent adversary would do. Then fix what you found.
Learns from completed work by comparing outcomes to expectations, identifying what worked and what didn't, extracting transferable lessons, and updating beliefs.
Assess compliance with applicable regulations by mapping requirements to current practices, identifying gaps, and prioritizing remediation.
Evaluate claims about how things relate. Verify the relationship type, test its strength, and identify its boundaries.
Handle relationship-related goals — clarify desired state, assess current reality, identify gaps, plan actions within your control.
Orderings for managing uncertainty — preserve options, handle irreversible decisions carefully, minimize downside.
Recognize when a mental model is fundamentally wrong, dismantle it, and build or adopt a replacement without losing momentum.
Reorder an existing list using an explicit objective, multi-dimensional scoring, dependency awareness, and sanity checks — producing a defensible best-to-worst order with confidence flags.
Find the highest-return action in any set of options. Lists options, estimates costs and benefits, calculates ROI, sensitivity-tests top candidates, and recommends the best path.
Handler for goals focused on REBUILDING from damage, not improving from baseline.
Meta-procedure for improving reasoning quality. Identifies the reasoning type in use, checks for common errors, and verifies logical structure.
Orderings for maximizing throughput, minimizing waste, and efficient resource utilization.
Identify and evaluate rights in a situation by mapping all parties, determining applicable rights, assessing respect, and identifying remedies.
Bridge the gap between thinking/planning and physical execution. Converts plans into actionable checklists with logistics, failure modes, and contingencies.
Run the detector suite against a target artifact to find ambiguity, proxy substitution, bundled checks, and more.
Classify utterances in a gate/procedure/document so the system doesn't answer the wrong thing.