kkemple

16 mods across 1 repository, 2 stars between them.

blockkit-builder

01

kkemple/skills

Skill Claude CodeCodex

Build and validate Slack Block Kit JSON using orchestrated multi-role adherence-coherence convergence. Use this skill when the user wants to build Block Kit layouts, generate block structures, create element compositions, construct surface-specific UIs, validate existing Block Kit JSON, or improve Block Kit payloads.…

2 4mo ago A 127 tokens original MIT

fixer

02

kkemple/skills

Agent

Surgical correction. The only agent that touches the artifact.

2 4mo ago A 0 tokens original MIT

generator

03

kkemple/skills

Agent

Discovery, planning, and artifact production. The first agent launched. Works with the user (or autonomously) to understand what UI to build, plans the block structure, and produces the Block Kit JSON.

2 4mo ago A 0 tokens original MIT

optimizer

04

kkemple/skills

Agent

Coherence and fitness. Domain and context lens.

2 4mo ago A 0 tokens original MIT

orchestrator

05

kkemple/skills

Agent

Orchestration, adjudication, state management.

2 4mo ago A 0 tokens original MIT

validator

06

kkemple/skills

Agent

Agent "validator" from kkemple/skills, covering validator, role, responsibilities, what you see and how to sweep.

2 4mo ago A 0 tokens original MIT

claude-tutor

07

kkemple/skills

Skill Claude CodeCodex

Produce a learner-facing study guide on a single concept using the Understanding by Design framework. Use this skill whenever a user wants to learn something specific — phrases like 'teach me X', 'help me understand Y', 'I want to learn Z', 'I'm trying to wrap my head around', 'walk me through', 'tutor me on', 'I need…

2 4mo ago A 168 tokens original MIT

complex-workflow

08

kkemple/skills

Skill Claude CodeCodex

Plan and execute complex multi-step tasks using a five-role agent system (Orchestrator, Planner, Auditor, Executor, Drift Monitor) with strict role isolation. Use this skill when the user has a complex, multi-step task that benefits from upfront planning and verified execution — refactors spanning multiple files…

2 4mo ago A 134 tokens original MIT

auditor

09

kkemple/skills

Agent

Judgment. Evaluates every step twice — before and after execution.

2 4mo ago A 0 tokens original MIT

drift-monitor

10

kkemple/skills

Agent

Drift detection. Measures semantic distance between the approved plan and the Orchestrator's current working state.

2 4mo ago A 0 tokens original MIT

executor

11

kkemple/skills

Agent

Execution. The only agent that performs the work.

2 4mo ago A 0 tokens original MIT

planner

12

kkemple/skills

Agent

Plan creation. User-facing during Phase 1 only.

2 4mo ago A 0 tokens original MIT

kkemple/skills

Skill Claude CodeCodex

Create a self-correcting validation skill that iteratively improves an artifact through parallel validity and fitness checks. Use this skill when the user has a recurring artifact type — code, schemas, configs, documents, designs — that needs both structural rule compliance AND quality assessment for an audience.…

2 4mo ago A 147 tokens original MIT

kkemple/skills

Skill Claude CodeCodex

Template for orchestrated multi-role adherence-coherence convergence. Copy and customize for a specific domain before use. This template implements a self-correcting architecture where an orchestrator orchestrates parallel validator (adherence) and optimizer (coherence) passes, merges their reports, and dispatches a…

2 4mo ago A 85 tokens original MIT

math-verify

15

kkemple/skills

Skill Claude CodeCodex

Verify mathematical correctness of a drafted mathematical paper using orchestrated multi-role adherence-coherence convergence. Use this skill when the user wants to check proofs, validate logical steps, verify invoked results, audit notation consistency, or confirm that theorem hypotheses are necessary and sufficient.…

2 4mo ago A 101 tokens original MIT

kkemple/skills

Skill Claude CodeCodex

Scan a project to identify artifacts that would benefit from self-correcting validation — where both structural rules and quality for an audience matter. Use this skill when the user wants to understand where automated review would help, is setting up quality processes, onboarding onto a new project, or asks things…

2 4mo ago A 110 tokens original MIT