A code-review command that examines code quality and suggests improvements. It can also challenge agreeable conclusions and examine wider effects of proposed changes.

Use it after coding when you want an evaluation of quality, security, accessibility, performance, or other review concerns. Type EVAL or EVALUATE to start it.

It helps find weaknesses and missed consequences after implementation, so improvement does not depend only on the developer’s first review.

eval — Command by JeremyDev87

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods

npx agentmods add commands/jeremydev87/codingbuddy/eval

Clone the repo

git clone --depth 1 https://github.com/JeremyDev87/codingbuddy

Per session 0 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.

When invoked 3,237 The whole file, excluding the scripts and references it only reads on demand.

Security scan A 0 findings. Scan, not verified.

Origin original No closer match found in the catalogue.

Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

Model	Per session	Once invoked
Fable 5	$0.00000	$0.03237
Opus 5	$0.00000	$0.01618
Sonnet 5	$0.00000	$0.00647
Haiku 4.5	$0.00000	$0.00324

Measured yesterday against content hash b9b70f590b77, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

eval scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured yesterday.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

packages/claude-code-plugin/commands/eval.md · 342 lines

How it starts

The opening of the file, as written. The whole thing — 342 lines — stays where its author put it; the contents beside it link to each section on GitHub.

EVAL Mode

Purpose

Evaluate code quality and suggest improvements

Mode Availability

Standalone Mode (no MCP server)

This command works without the CodingBuddy MCP server. In standalone mode, you get:

Mode detection and transition
Code quality evaluation
Anti-sycophancy analysis
Devil's Advocate and Impact Radius analysis

MCP Enhanced Mode

With the CodingBuddy MCP server connected, you additionally get:

parse_mode with full context tracking
Specialist agent dispatch for comprehensive evaluation
Contextual checklists (security, accessibility, performance)
Cross-session context persistence

Activation

Invoke via /codingbuddy:eval (namespaced) or the EVAL keyword.

Legacy bare /eval is deprecated. See Migration Guide.

This command activates EVAL mode for the CodingBuddy workflow.

Instructions

Important:

EVAL mode is not automatic after ACT
User must explicitly request EVAL by typing EVAL
Default behavior after ACT: Return to PLAN (without evaluation)
Use EVAL when you want iterative improvement and refinement

Trigger:

Type EVAL after completing ACT
Type EVALUATE (also accepted)
Korean: 평가해 or 개선안 제시해

🔴 Agent Activation (STRICT):

When EVAL is triggered, Code Reviewer Agent (loaded via parse_mode in MCP mode; defaults in standalone mode) MUST be automatically activated
The Agent's evaluation framework and all mandatory requirements MUST be followed
See agent documentation for complete evaluation framework

Purpose: Self-improvement through iterative refinement

What EVAL does (with Code Reviewer Agent):

Analyze Implementation (via Code Reviewer Agent)
- Review what was done in ACT
- Check adherence to project rules
- Verify quality standards met
- 🔴 Required: Follow Code Reviewer Agent's evaluation framework
Assess Quality (via Code Reviewer Agent mandatory perspectives)
- 🔴 Code quality (SOLID, DRY, complexity)
  - Required: When evaluating code quality, reference Code Quality Specialist Agent (dispatched via parse_mode in MCP mode) modes.evaluation framework for SOLID principles, DRY, complexity analysis, and design patterns assessment
- 🔴 Architecture (layer boundaries, dependency direction, type safety)
  - Required: When evaluating architecture, reference Architecture Specialist Agent (dispatched via parse_mode in MCP mode) framework for layer boundaries, dependency direction, and type safety assessment
- 🔴 Test coverage (90%+ goal)
  - Required: When evaluating tests, reference Test Strategy Specialist Agent (dispatched via parse_mode in MCP mode) modes.evaluation framework for test coverage, TDD workflow, and test quality assessment
- 🔴 Performance (build size, execution optimization)
  - Required: When evaluating performance, reference Performance Specialist Agent (dispatched via parse_mode in MCP mode) framework for build size, execution optimization, and performance metrics assessment
- 🔴 Security (XSS/CSRF, authentication/authorization)
  - Required: When evaluating security, reference Security Specialist Agent (dispatched via parse_mode in MCP mode) framework for OAuth 2.0, JWT, CSRF/XSS protection assessment
- 🔴 Accessibility (WCAG 2.1 AA compliance)
  - Required: When evaluating accessibility, reference Accessibility Specialist Agent (dispatched via parse_mode in MCP mode) framework for WCAG 2.1 AA compliance verification
- 🔴 SEO (metadata, structured data)
  - Required: When evaluating SEO, reference SEO Specialist Agent (dispatched via parse_mode in MCP mode) framework for metadata, structured data, and search engine optimization assessment
- 🔴 UI/UX Design (visual hierarchy, UX patterns)
  - Required: When evaluating UI/UX design, reference UI/UX Designer Agent (dispatched via parse_mode in MCP mode) framework for visual hierarchy, UX laws, and interaction patterns assessment
- 🔴 Documentation Quality (documentation, cursor rules, AI prompts)
  - Required: When evaluating documentation, cursor rules, or AI prompts, reference Documentation Specialist Agent (dispatched via parse_mode in MCP mode) modes.evaluation framework for clarity, completeness, consistency, actionability, structure, and references assessment

Read the full file on GitHub · 342 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

yesterday First seen · 342 lines · 0 tokens per session scan A b9b70f590b77

Subscribe to this mod's changes

eval is a command published in the GitHub repository JeremyDev87/codingbuddy (31 stars, last pushed 4mo ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 3,237 tokens. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.

Code review Languages Security ai-agents ai-coding ai-coding-assistant ai-rules claude-code typescript

eval

Getting it into your agent

What it costs to keep this loaded

Grade A, and why

Nothing flagged

How it starts

EVAL Mode

Purpose

Mode Availability

Standalone Mode (no MCP server)

MCP Enhanced Mode

Activation

Instructions

What this file has done since we first saw it

Other commands, from other repositories

extract

codex-review

raw-export

stats

update-plugins

setup