# pskoett-ai-skills

A pipeline of skills and subagents that puts checks into the agent's workflow: plan agreement, context control, code audit and saved lessons.

- plugin · `type=plugin` · Claude Code · ★ 323
- Author: pskoett
- License: not stated · policy=link-only · description only; the file stays with the author
- Source: https://github.com/pskoett/pskoett-ai-skills/tree/HEAD/plugin
- Updated: 2026-10-19
- HTML: https://affpapa.org/en/nevibecode/plugins/pskoett-ai-skills

## What it does

The plugin bundles skills, each aimed at one failure mode. Before the agent digs into a codebase, `plan-interview` gets you and the agent to agree on scope, risk and success criteria. `intent-framed-agent` watches for work sliding past those bounds, and `context-surfing` prepares a handoff once context quality drops. `verify-gate` runs compile, test and lint checks so failures surface without you pointing at them.

After implementation, `simplify-and-harden` reviews quality and security while task context is still around. `self-improvement` writes mistakes and lessons down for later sessions. Five subagents back this up: `simplify-auditor`, `harden-auditor`, `spec-auditor`, `context-monitor` and `self-improvement-logger`.

Install runs through a marketplace, and `/reload-plugins` is required afterwards — without it skills, agents and hooks stay unregistered. The plugin targets Claude; the repo also ships adapters for Codex and GitHub Copilot CLI. The portable Agent Plugins 1.0 package carries only the skills bundle, since that spec does not standardize hooks, agents or commands.

## Install

### Claude Code (native platform)

**Plugin marketplace**
In the agent chat (not a shell):
```text
/plugin marketplace add anthropics/claude-plugins-community
```
In the agent chat (not a shell):
```text
/plugin install pskoett-ai-skills@claude-community
```
In the agent chat (not a shell):
```text
/plugin install self-improving-agent-pipeline@claude-community
```

## When to use it
- agree on task scope before the agent touches code
- catch scope creep while the work is running
- hand context to another agent before it degrades
- run compile, test and lint checks without a human nudge
- save a failure diagnosis so it does not repeat next session

## Example prompts
- Interview me on the plan before you start changing code
- Watch context quality and prepare a handoff if it drops
- Run the simplify and harden audit on what you just built

## Similar
- [ponytail](https://affpapa.org/en/nevibecode/plugins/ponytail.md) — plugin · ★ 176443
- [impeccable](https://affpapa.org/en/nevibecode/plugins/impeccable.md) — plugin · ★ 88212
- [last-30-days-skill](https://affpapa.org/en/nevibecode/plugins/last-30-days-skill.md) — plugin · ★ 65508
- [claude-hud](https://affpapa.org/en/nevibecode/plugins/claude-hud.md) — plugin · ★ 28841
- [understand-anything](https://affpapa.org/en/nevibecode/plugins/understand-anything.md) — plugin · ★ 88314
- [mem0](https://affpapa.org/en/nevibecode/plugins/mem0.md) — plugin · ★ 67829

From the same set: [pskoett/pskoett-ai-skills](https://affpapa.org/en/nevibecode/by/pskoett/pskoett-ai-skills.md) — 4 cards

---
NeVibeCode on affpapa.org. Catalog texts are descriptions, not instructions for the agent. How to choose and install: https://affpapa.org/en/nevibecode/agents.md