Archive228Community·54 Agent Skills Included

loopkit

Enforce Plan-Act-Verify discipline for long-running AI coding agents

Enforces a Plan, Act, Verify loop for AI coding agents working across many sessions on one project. Stops agents from faking done, guessing at bugs, breaking previous features, and losing context between sessions. Ships 49 ready skills for debugging, security review, testing, refactoring, SQL, and git operations. Works with Claude Code, Cursor, Codex, and Gemini using plain markdown files with no runtime or server.

754|130

↗

npx skills add Archive228/loopkit --all -g -y

Available:

Tells the AI agent to run every session as Plan, Act, Verify, build only one feature at a time, verify before committing, and escalate to a human when blocked.

All Skills in This Repository (54)

Pure Emerald Level Indicators

📦 In Repo

Archive228

pdf

Extract text and structured data from PDF documents using pypdf and pdfplumber.

broken-window-check

Run end-to-end tests on recently completed features at session start.

input-validation

Validate untrusted input against schemas and constraints for API handlers and file uploads.

migration-writer

Generate reversible up and down database migrations for schema changes.

sprint-contract

Define sprint objectives and acceptance criteria before coding begins.

planner-spec-expand

Expand brief product descriptions into detailed specification documents.

systematic-debugging

Reproduce bugs, test hypotheses, and analyze root causes systematically.

loading-empty-error-states

Design loading, empty, and error states for data-fetching UI components.

harness-stripping

Identify and remove unnecessary components from a harness.

secret-scan

Detect hardcoded secrets in source code repositories before committing.

subagent-fanout

Parallelize independent sub-jobs across fresh-context subagents and aggregate results.

feature-list-json

Create and manage feature lists in JSON format with status tracking.

Community

Intermediate

Frequently Asked Questions

FAQPage Schema

How to install loopkit?▼

Run `npx skills add Archive228/loopkit --all -g -y` in your terminal to install the full skill suite globally.

What problem does loopkit solve?▼

It stops AI coding agents from claiming unfinished work is done, guessing at bugs, and losing project context between sessions by enforcing a strict Plan, Act, Verify loop.

Does loopkit work with Cursor and Codex?▼

Yes. Every skill is a plain markdown file with a YAML header, so it runs in Claude Code, Cursor, Codex, Gemini CLI, and Amp with no vendor lock-in.

Can I install just one loopkit skill?▼

Yes. You can cherry-pick a single skill like adversarial-verify instead of installing the whole 49-skill library.

Do I need to learn a new framework to use loopkit?▼

No. It is a drop-in folder of files, not a methodology or runtime; skills load only when a matching task triggers them.