name	verification-and-reliability
description	Use first when the user asks for tests, correctness, high reliability, static reasoning, proof, or safety evidence.

Verification and Reliability

Use first when the user asks for tests, correctness, high reliability, static reasoning, proof, or safety evidence. This is a router skill: use it to select the smallest relevant leaf skill, then read that leaf skill before doing the work.

Route First

Choose the proof style before changing code: test-first behavior discovery, specification-driven prevention, or sound approximation.

Restate the user's task as one concrete force or uncertainty.
Choose the first route that directly names that force.
Read the selected linked leaf SKILL.md before implementing, reviewing, or advising.
Load a second leaf only when the task has two independent forces that both affect the outcome.
If no route fits, continue without a leaf skill and say the category did not match.

Routes

Leaf Skill	Use When
`test-driven-development`	Use when the next observable behavior can be captured by a failing test before implementation.
`cleanroom-software-engineering`	Use when reliability risk requires precise specification, defect prevention, and small verifiable increments.
`abstract-interpretation`	Use when many possible executions must be reasoned about with a sound over-approximation.
`design-by-contract`	Use when reliability depends on explicit preconditions, postconditions, or invariants at a boundary.
`stepwise-refinement`	Use when proof depends on preserving a high-level specification through implementation steps.

Avoid

If the behavior is still ambiguous, route to ambiguity-and-learning before writing tests.
If the issue is primarily human workflow evidence, route to workflow-and-operations.

Prompting Pattern

Before loading a leaf, answer briefly:

Category force: What makes this task belong here?
Chosen route: Which leaf skill most directly matches the force?
Why not others: Which nearby route was rejected and why?

Then load the chosen leaf skill and follow its workflow. Do not blend every nearby theory into the task; route narrowly and let evidence pull in more context only when needed.

Más de este repositorio

mismo repositorio

skill-router

AnamKwon/agent-skill-router

Use when organizing a large local skill library into category router skills, refreshing a compact skill inventory, or deciding which leaf skill bodies need to be read before agent classification.

2026-06-189

ambiguity-and-learning

AnamKwon/agent-skill-router

Use first for ambiguous failures, messy project context, surprising results, repeated failures, postmortems, or contested product and architecture framing.

2026-06-189

boundary-and-modeling

AnamKwon/agent-skill-router

Use first for API boundaries, module ownership, domain modeling, contracts, architecture seams, and recurring design structures.

2026-06-189

code-change-core

AnamKwon/agent-skill-router

Use first for bounded code edits, refactors, implementation, simplification, control-flow cleanup, or maintainer clarity before choosing a more specific theory skill.

2026-06-189

skill-router

AnamKwon/agent-skill-router

Use to organize a large skill library into category routers, route new skills, and propose new categories when existing routers are insufficient.

2026-06-189

workflow-and-operations

AnamKwon/agent-skill-router

Use first for work shaped by users, tools, procedures, runtime evidence, operational artifacts, interfaces, or organizational workflows.

2026-06-189

Verification and Reliability

Route First

Choose the proof style before changing code: test-first behavior discovery, specification-driven prevention, or sound approximation.

Restate the user's task as one concrete force or uncertainty.

Choose the first route that directly names that force.

Read the selected linked leaf SKILL.md before implementing, reviewing, or advising.

Load a second leaf only when the task has two independent forces that both affect the outcome.

If no route fits, continue without a leaf skill and say the category did not match.

Routes

Leaf Skill

Use When

test-driven-development

Use when the next observable behavior can be captured by a failing test before implementation.

cleanroom-software-engineering

Use when reliability risk requires precise specification, defect prevention, and small verifiable increments.

abstract-interpretation

Use when many possible executions must be reasoned about with a sound over-approximation.

design-by-contract

Use when reliability depends on explicit preconditions, postconditions, or invariants at a boundary.

stepwise-refinement

Use when proof depends on preserving a high-level specification through implementation steps.

Avoid

If the behavior is still ambiguous, route to ambiguity-and-learning before writing tests.

If the issue is primarily human workflow evidence, route to workflow-and-operations.

Prompting Pattern

Before loading a leaf, answer briefly:

Category force: What makes this task belong here?

Chosen route: Which leaf skill most directly matches the force?

Why not others: Which nearby route was rejected and why?

Then load the chosen leaf skill and follow its workflow. Do not blend every nearby theory into the task; route narrowly and let evidence pull in more context only when needed.