Skip to content

nikitakurpas/standards

v0.1.3MIT

Engineering principles as skills. /with-standards applies them to the current task.

blast-radius

Finds what a change could break beyond its diff, and proves the one fact that makes it safe by running real code.

in-your-own-words

Restates the user's goals and the problem they're solving, so you can check the agent understood.

ponytail-docs

Minimize project documentation without losing truth. Use when writing or reviewing repository docs, or when asked to prevent documentation sprawl. Not for general prose.

principle-boundary-discipline

Concentrates validation and error handling at system boundaries, trusts internal types, and keeps business logic pure.

principle-build-the-lever

Builds the script, codemod, or generator that does or proves the work, so a reviewer can rerun it.

principle-encode-lessons-in-structure

Turns a repeated instruction or correction into a lint, metadata flag, runtime check, or script.

principle-exhaust-the-design-space

Builds 2-3 competing prototypes for a decision with no precedent and compares them before committing.

principle-experience-first

Chooses user delight over implementation convenience, and fewer polished features over more rough ones.

principle-fix-root-causes

Traces each symptom to the first place the system went wrong and fixes it there.

principle-foundational-thinking

Settles core types, data structures, and shared state before writing logic.

principle-guard-the-context-window

Keeps the main context lean by routing bulk reading to subagents and keeping only their summaries.

principle-laziness-protocol

Prefers deletion and the smallest change that solves the problem over new abstractions or layers.

principle-make-operations-idempotent

Designs commands, lifecycle steps, and loops to reach the same end state after crashes, restarts, and retries.

principle-migrate-callers-then-delete-legacy-apis

Moves every caller to a new internal API and deletes the old one in the same wave.

principle-minimize-reader-load

Cuts the layers and hidden state a reader must trace, collapsing one-caller wrappers and shrinking mutable scope.

principle-model-the-domain

Encodes domain rules in one structure, such as a state machine, typed model, or table, instead of scattered conditionals.

principle-never-block-on-the-human

Proceeds on reversible work instead of asking permission, then presents the result for correction.

principle-outcome-oriented-execution

Drives rewrites and migrations straight to the target architecture, without throwaway compatibility code.

principle-prove-it-works

Verifies finished work against the real artifact, not a proxy, a self-report, or a passing build.

principle-redesign-from-first-principles

Integrates a new requirement by redesigning as if it had been there from the start.

principle-sequence-verifiable-units

Splits multi-step work into small units that each end in a check, and orders commits and PRs so the sequence proves itself.

principle-subtract-before-you-add

Removes dead code and redundant pieces before adding, refactoring, or rewriting.

principle-test-behavior-not-implementation

Writes tests that assert observable behavior, so they fail when behavior breaks and survive internal changes.

principle-type-system-discipline

Designs types that make illegal states unrepresentable, brand primitives, and parse external data at boundaries.

technical-writing

Writing standard for docs, readmes, RFCs, PR descriptions, and commit messages, built on Diátaxis, Google developer style, STE, and Global English.

typescript-best-practices

TypeScript rules for modeling types, narrowing safely, validating at boundaries, and testing real code.

unslop

Removes AI writing patterns such as filler, hedging, stock vocabulary, and jargon.

with-standards

Applies the Standards principles to the current task and names the ones that shaped each decision. Use when writing, refactoring, or reviewing code or docs, or designing a change.