Zhaono1 Agent Playbook - Setup Guide¶
Source: zhaono1/agent-playbook GitHub: zhaono1/agent-playbook Skills: 24 skills · 50.4K total installs Category: Agent Workflows First Seen: catalogued August 17, 2026 evening sweep (self-improving-agent on skills.sh since January 22, 2026; queue re-verification promoted the cluster from stale API numbers) Quality Tier: 🟡 Trusted - flagship self-improving-agent carries Gen Agent Trust Hub Warn and Socket Warn (Snyk Pass); both named in Limitations
A 24-role agent lifecycle suite: self-improving-agent (33.1K installs) implements a multi-memory feedback loop with hooks-based self-correction, and the rest of the pack covers the full build cycle from PRD planning through code review, testing, and deployment. The cluster had been parked on stale API-sum numbers (top skill presumed under 1.2K); its publisher page reads 50.4K total.
Installation¶
npx skills add zhaono1/agent-playbook
Individual skills install with the explicit repo form:
npx skills add https://github.com/zhaono1/agent-playbook --skill self-improving-agent
Prerequisites¶
| Requirement | Details |
|---|---|
| Node.js + npx | For the skill installer |
| Agent runtime | Claude Code, Cursor, or any skills-compatible agent |
What It Provides¶
| Skill | Installs | Purpose |
|---|---|---|
| self-improving-agent | 33.1K | Multi-memory (semantic + episodic + working) self-correction loop with hooks (before_start, after_complete, on_error) and evolution markers |
| planning-with-files | 1.1K | File-backed planning discipline |
| security-auditor | 1.0K | Security review of proposed changes |
| architecting-solutions | 851 | Solution architecture guidance |
| skill-router | 839 | Routing between skills by task |
| prd-planner / prd-implementation-precheck | 807 / 690 | PRD planning and pre-implementation checks |
| workflow-orchestrator | 798 | Multi-step workflow coordination |
| test-automator / qa-expert / debugger | 785 / 728 / 728 | Test automation, QA, and debugging roles |
| figma-designer | 784 | Design handoff interpretation |
| code-reviewer / refactoring-specialist | 723 / 716 | Review and refactoring roles |
| session-logger | 724 | Session capture for continuity |
| api-designer / api-documenter / documentation-engineer | 681 / 691 / 689 | API and docs roles |
| performance-engineer | 682 | Performance analysis |
| auto-trigger / commit-helper / create-pr / deployment-engineer / long-task-coordinator | 648-669 | Trigger automation, commit, PR, deploy, and long-task roles |
Quick Start¶
- Install:
npx skills add zhaono1/agent-playbook - Ask: "set up the self-improving loop for this project's skills"
- The skill wires hooks on skill events and logs evolution markers for every change
CorpusIQ Use Cases¶
| Use Case | How |
|---|---|
| Self-evolution | The multi-memory feedback loop maps to our own skill-patching discipline |
| Lifecycle coverage | PRD planner through deployment engineer gives a full pipeline in one pack |
| Session continuity | session-logger and planning-with-files reinforce our handoff-page pattern |
Integration with CorpusIQ¶
CorpusIQ is the connector layer that lets AI assistants read the systems a business already runs on - accounting, payments, CRM, email, files and analytics - and answer with the source named. It is read-only by design: it retrieves data and never writes to, or acts in, a connected system, which is what makes it safe to point at live accounts.
This page is part of the CorpusIQ directory, the reference we maintain for the MCP server and agent-skill ecosystem. Run alongside CorpusIQ, a third-party server like this one can be combined with a company's own business data in a single answer.
Limitations / Verification¶
- Security audits on self-improving-agent: Gen Agent Trust Hub Warn, Socket Warn, Snyk Pass - trusted with both warns named; the skill writes agent memory files, which drives the audit warnings
- Publisher-page total verified (50.4K across 24 skills); repo at 73 GitHub stars
- A sibling publisher's self-improving-agent is already guided separately (charon-fan/agent-playbook at 32.2K); this guide covers zhaono1's full 24-skill suite on its own publisher-page numbers
- 23 of 24 skills sit under 1.2K installs - the suite's weight is concentrated in self-improving-agent
npx skills add zhaono1/agent-playbook # verify install works
Related¶
Powered by CorpusIQ