Agent Design Review

Review an LLM agent design and find where it will be unreliable, expensive, or unsafe. Use when asked to review an agent architecture, critique a multi-step/tool-using agent, debug an agent that loops or goes off-task, or harden an agent before launch. Produces a structured review — task fit, control flow, tools, memory/context, failure handling, cost, and safety — with prioritised findings and fixes.

1,573 stars 281 forksAI, LLMs & Agents Plugin marketplace

Install

Claude Code plugin
/plugin marketplace add mohitagw15856/pm-claude-skills

The repo is a Claude Code plugin marketplace. Run /plugin afterwards to pick the plugin that includes this skill.

skills CLI (Claude Code, Cursor, Codex and more)
npx skills add mohitagw15856/pm-claude-skills --skill agent-design-review
Manual install
git clone --depth 1 https://github.com/mohitagw15856/pm-claude-skills.git /tmp/pm-claude-skills
mkdir -p ~/.claude/skills
cp -r /tmp/pm-claude-skills/plugins/pm-agentops/skills/agent-design-review ~/.claude/skills/agent-design-review

Personal skills live in ~/.claude/skills. Use .claude/skills in a project folder to share a skill with your team.

From the SKILL.md

# Agent Design Review Skill

Most agents don't fail because the model is weak — they fail because the *design* lets them loop, call the
wrong tool, lose the thread across steps, or burn tokens with no stopping rule. This skill reviews an agent's
architecture against the decisions that actually determine reliability, and ranks the fixes — so "it works in
the demo but not in prod" becomes a specific list of changes. (Writing a new agent spec? Use
[`agent-spec`](../agent-spec/SKILL.md).)

## Working from a brief

Given a sketch ("a research agent that searches, reads, and writes a report"), **deliver the full review
anyway** — infer the likely control flow and tools, label the inference, and flag what to confirm. Never
withhold the review for missing detail.

## Required Inputs

Ask for these only if they aren't already provided (else infer and label):

- **What the agent does** — its goal, and what a successful run produces.
- **Control flow** — single prompt, plan-then-execute, ReAct loop, or multi-agent; and the stopping condition.
- **Tools & actions** — what it can call, and which actions have side effects (write, send, pay).
- **Memory & context** — what state carries across steps, and how context is kept in budget.
- **Constraints** — latency, cost per run, and the trust boundary (untrusted input? real-world actions?).

## Output Format

### Agent Review: [agent]

**1. Summary** — will this be reliable in production? The top 3 risks and the single change that helps most.
Read the full skill on GitHub

Details

Skill name
agent-design-review
Path
plugins/pm-agentops/skills/agent-design-review/SKILL.md
License
MIT
Repo updated
18 November 2026

FAQ

What does the Agent Design Review skill do?

Review an LLM agent design and find where it will be unreliable, expensive, or unsafe. Use when asked to review an agent architecture, critique a multi-step/tool-using agent, debug an agent that loops or goes off-task, or harden an agent before launch. Produces a structured review — task fit, control flow, tools, memory/context, failure handling, cost, and safety — with prioritised findings and fixes.

How do I install the Agent Design Review skill?

Claude Code plugin: /plugin marketplace add mohitagw15856/pm-claude-skills. Skills go in ~/.claude/skills (personal) or .claude/skills (project), and Claude loads them automatically when a task matches the description.

Is Agent Design Review popular?

It lives in mohitagw15856/pm-claude-skills, which has 1,573 GitHub stars and 281 forks, and is included in 1 curated skill list.

Related skills

Ab Test Planner

1.6k

mohitagw15856/pm-claude-skills

Diseña pruebas A/B estadísticamente rigurosas para características de producto, cambios de UI, flujos de onboarding y experimentos de precios. Úsalo cuando te pidan configurar un experimento, diseñar una prueba A/B, calcular tamaño de muestra o interpretar resultados de pruebas. Produce un plan de prueba completo con hipótesis, definición de variantes, tamaño de muestra, estimación de duración, métricas de seguridad y una guía de interpretación de resultados.

Testing & QAIn 1 list281

Ab Test Planner

1.6k

mohitagw15856/pm-claude-skills

प्रोडक्ट फ़ीचर, UI बदलाव, ऑनबोर्डिंग फ़्लो और प्राइसिंग प्रयोगों के लिए सांख्यिकीय रूप से कठोर A/B टेस्ट डिज़ाइन करें। जब कोई प्रयोग सेट करने, A/B टेस्ट डिज़ाइन करने, सैंपल साइज़ निकालने या टेस्ट के नतीजे समझाने को कहे, तब इसका उपयोग करें। परिकल्पना, वेरिएंट की परिभाषा, सैंपल साइज़, अवधि का अनुमान, गार्डरेल मेट्रिक और नतीजों की व्याख्या की गाइड के साथ एक पूरा टेस्ट प्लान तैयार करता है।

Testing & QAIn 1 list281

Accessible Travel Planner

1.6k

mohitagw15856/pm-claude-skills

Summarise what Accessible Travel Planner does in one line. Use when asked to [trigger phrases the user would say]. Produces [the concrete artifact].

Frontend & DesignIn 1 list281

Agent Hiring Panel

1.6k

mohitagw15856/pm-claude-skills

Summarise what Agent Hiring Panel does in one line. Use when asked to [trigger phrases the user would say]. Produces [the concrete artifact].

AI, LLMs & AgentsIn 1 list281

Prompt Engineer

2k

rmyndharis/antigravity-skills

Expert prompt engineer specializing in advanced prompting techniques, LLM optimization, and AI system design. Masters chain-of-thought, constitutional AI, and production prompt strategies. Use when building AI features, improving agent performance, or crafting system prompts.

AI, LLMs & Agents295

Agent Incident Postmortem

1.6k

mohitagw15856/pm-claude-skills

Run a blameless postmortem for an incident caused by an AI agent or LLM feature — hallucinated facts shipped to users, runaway tool use, prompt injection, cost blowouts, or wrong actions taken autonomously. Use when asked to write up an AI incident, analyse why an agent did something wrong, or produce corrective actions after an LLM failure. Produces a structured postmortem with trace reconstruction, a root-cause layer analysis, and corrective actions including a permanent regression case. For non-AI production incidents use incident-postmortem.

AI, LLMs & AgentsIn 1 list281

Browse the most popular Claude skills, all ai, llms & agents skills, or write your own with the skills generator.