Caesar Harness Agent

The definitive collection of harness subagents, specialized AI assistants designed for specific development tasks and native cross-platform coding agent outputs

Write expert coding subagents once. Emit native agent outputs everywhere.

Caesar-Harness-Agent serves as the definitive collection of harness subagents, specialized AI assistants designed for specific development tasks and native cross-platform coding agent outputs.

Why Caesar Harness Agent

Every AI coding tool invented its own agent file format. Maintaining the same backend reviewer, Rust expert, or security auditor prompt across multiple tools by hand creates drift and stale behavior.

Caesar Harness Agent keeps one canonical Markdown + YAML definition per expert agent, then transpiles it into each tool's native format with schema validation on every build. One source of truth. No per-tool copy-paste. No prompt drift.

Installation Methods

Method 1: Automatic via Caesar CLI (Recommended)

Add agent plugins directly using npx @caesar/cli:

# Add to current project (all supported tools)
npx @caesar/cli add @caesar/lang

# Add for specific tools only
npx @caesar/cli add @caesar/data-ai --tool claude --tool opencode

# Add globally (for tools that support global agent directories)
npx @caesar/cli add @caesar/core-dev --global

Manage your installed plugins:

npx @caesar/cli list
npx @caesar/cli list --global
npx @caesar/cli remove @caesar/lang

Plugins support local paths, npm packages, and GitHub repositories:

npx @caesar/cli add github:user/custom-agents

Method 2: Project devDependency & Version Pinning

Manage plugin versions within your project's package lifecycle:

1. Install the category plugin:

   pnpm add -D @caesar/lang
   # or npm install --save-dev @caesar/lang

2. Run the local CLI to link artifacts:

   npx caesar add @caesar/lang

Method 3: Automated via postinstall Hook

To ensure all developers and CI pipelines have the correct agent configs synced: Add this to your project's package.json:

{
  "scripts": {
    "postinstall": "caesar add @caesar/lang"
  }
}

Method 4: Manual Copy-Paste (Direct Folder Mapping)

If you prefer not to use the CLI, you can install the npm package or clone the repository, then copy files manually:

Tool	Source Path (in @caesar/{category}/dist/)	Target Path (Project Root)
Claude Code	claude/.claude/agents/*.md	.claude/agents/
OpenCode	opencode/.opencode/agents/*.md	.opencode/agents/
Kiro	kiro/.kiro/agents/*.json	.kiro/agents/
Codex	codex/.codex/agents/*.toml	.codex/agents/
Factory / Droid	factory/.factory/droids/*.md	.factory/droids/
GitHub Copilot	copilot/.github/agents/*.agent.md	.github/agents/
Gemini CLI	gemini/.gemini/agents/*.md	.gemini/agents/
OpenHands	openhands/.agents/skills/*	.agents/skills/
Kilo Code	kilo/.kilocodemodes	.kilocodemodes
Roo Code	roo/.roomodes	.roomodes
Cursor	cursor/.cursor/rules/*.mdc	.cursor/rules/
Claude Marketplace	claude-plugin/.claude-plugin/marketplace.json, plugin.json and claude-agents/	.claude-plugin/ and claude-agents/
Fallbacks	agents-md/AGENTS.md	AGENTS.md

Interactive Auto-Detection

When you run caesar add <source> without specifying the --tool flag, the CLI automatically detects which AI coding tools are configured on your machine.

• Interactive Terminal (TTY): The CLI will scan your workspace and global directories, and ask you:

  Detected active tools: claude, opencode. Install for these? (Y/n): 
  

  If you accept, it installs only for those tools. If you decline, it will let you enter a custom comma-separated list of tools.
• Non-Interactive (CI/CD / Postinstall): The CLI will automatically detect active tools and install only for them. If no active tools are found, it acts as a safe no-op to prevent writing unwanted folders and polluting your workspace.

Native agent outputs

Tool	Output	Tier
Claude Code	.claude/agents/*.md	native
OpenCode	.opencode/agents/*.md	native
Kiro	.kiro/agents/*.json	native
Codex	.codex/agents/*.toml	native
Factory / Droid	.factory/droids/*.md	native
GitHub Copilot / VS Code	.github/agents/*.agent.md	native
Gemini CLI	.gemini/agents/*.md	native, inherit-only emit
OpenHands	.agents/skills/*/SKILL.md	native skill output
Kilo Code	.kilocodemodes	native aggregate output
Roo Code	.roomodes	native aggregate output
Cursor	.cursor/rules/*.mdc	native per-agent MDC
Claude Marketplace	.claude-plugin/marketplace.json, .claude-plugin/plugin.json, claude-agents/*.md	native aggregate output
Cursor, Windsurf, Cline, Antigravity, Amp	shared AGENTS.md routing index	fallback

Full matrix and per-tool limits: docs/tool-support-matrix.md.

Target Tools Configuration & Usage

Once the agents are copied or installed into their respective target paths, here is how you invoke and use them in each supported tool.

1. Claude Code

• Path: Local .claude/agents/ or Global ~/.claude/agents/
• Installation:

  claude plugin marketplace add Caesar-Nexus-Labs/caesar-harness-agent
  claude plugin install <plugin-name>

• Usage: Run claude. Inside the terminal, type / for autocomplete or /<agent-slug> (e.g., /typescript-pro).

2. OpenCode

• Path: Local .opencode/agents/ or Global ~/.opencode/agents/
• Installation:

  opencode subagent add Caesar-Nexus-Labs/caesar-harness-agent

• Usage: Run opencode or open the editor interface. Select subagents from the session setup menu.

3. Kiro

• Path: Local .kiro/agents/ or Global ~/.kiro/agents/
• Installation:

  kiro agent install Caesar-Nexus-Labs/caesar-harness-agent

• Usage: Run kiro CLI or use the Kiro IDE/UI mode selector and specify the agent by name.

4. Factory / Droid

• Path: Local .factory/droids/ or Global ~/.factory/droids/
• Installation:

  factory agent add Caesar-Nexus-Labs/caesar-harness-agent

• Usage: Run factory CLI or access the web build dashboard. Select droids in the build execution interface.

5. GitHub Copilot / VS Code

• Path: Local .github/agents/
• Installation:

  copilot agent add Caesar-Nexus-Labs/caesar-harness-agent

• Usage: In VS Code Copilot Chat view, type @agent-slug (e.g., @typescript-pro) to route conversation.

6. OpenHands

• Path: Local .agents/skills/
• Installation:

  openhands agent install Caesar-Nexus-Labs/caesar-harness-agent

• Usage: Run OpenHands web IDE or its CLI tool. Agents automatically inject these folders as workspace skills.

7. Codex

• Path: Local .codex/agents/ or Global ~/.codex/agents/
• Installation:

  npx @caesar/cli add github:Caesar-Nexus-Labs/caesar-harness-agent --tool codex

• Usage: Run codex CLI or launch the Codex web IDE. It automatically scans .codex/agents/.

8. Gemini CLI & Antigravity (Agy)

• Path: Local .gemini/agents/ or Global ~/.gemini/agents/
• Installation:

  npx @caesar/cli add github:Caesar-Nexus-Labs/caesar-harness-agent --tool gemini

• Usage: Run gemini or antigravity. Target the mode using the mode picker or session arguments.

9. Kilo Code

• Path: Local .kilocodemodes (single YAML file)
• Installation:

  npx @caesar/cli add github:Caesar-Nexus-Labs/caesar-harness-agent --tool kilo

• Usage: Read from the UI dropdown in Kilo.

10. Roo Code

• Path: Local .roomodes (single YAML file)
• Installation:

  npx @caesar/cli add github:Caesar-Nexus-Labs/caesar-harness-agent --tool roo

• Usage: Select the custom mode from the mode dropdown in the VS Code sidebar.

11. Cursor (Native MDC Rules)

• Path: Local .cursor/rules/
• Installation:

  npx @caesar/cli add github:Caesar-Nexus-Labs/caesar-harness-agent --tool cursor

• Usage: Run cursor .. Mention rules in chat using @rule-name or rely on automatic glob matching.

12. Claude Plugin Marketplace

• Path: Local .claude-plugin/
• Installation:

  npx @caesar/cli add github:Caesar-Nexus-Labs/caesar-harness-agent --tool claude-marketplace

• Usage: Configures the local registry for custom plugin distributions. Run your local marketplace server to serve these to Claude.

13. Fallbacks (Windsurf, Cline, Amp)

• Path: Local AGENTS.md
• Installation:

  npx @caesar/cli add github:Caesar-Nexus-Labs/caesar-harness-agent --tool agents-md

• Usage: Fallback tools read AGENTS.md as shared instructions. Cline also checks .clinerules/.

Expert coding agents and prompt library

Caesar-Harness-Agent currently ships 151 expert coding agents across 11 categories.

#	Category	Count	Examples
01	core-development	7	backend-developer, api-designer, microservices-architect
02	language-specialists	28	typescript-pro, rust-engineer, python-pro, flutter-expert
03	infrastructure	18	kubernetes-specialist, terraform-engineer, sre-engineer
04	quality-security	15	code-reviewer, security-auditor, penetration-tester
05	data-ai	19	ai-engineer, ml-engineer, llm-architect, eval-engineer
06	developer-experience	12	refactoring-specialist, mcp-developer, build-engineer
07	specialized-domains	11	blockchain-developer, fintech-engineer, embedded-systems
08	business-product	8	product-manager, business-analyst, technical-writer
09	meta-orchestration	11	multi-agent-coordinator, context-manager, policy-guardrail-designer
10	research-analysis	6	research-analyst, search-specialist, first-principles-thinking
11	marketing	16	ad-copy-creator, campaign-architect, seo-content-strategist, social-media-planner

Each category can ship as its own npm package of prebuilt artifacts (@caesar/lang, @caesar/infra, and others) so projects install only the roles they need.

How it works

agents/{NN-category}/{agent}.md   canonical source (Markdown + YAML)
        │
        ▼  @caesar/agents-core  (schema → transpiler)
        │
        ▼  core/ → registry/ → adapters/ & serializers/
        │
        ▼  per-tool output validators → write-outputs
        │
        ▼
dist/{tool}/…   native files for each tool   +   dist/agents-md/AGENTS.md

The transpilation engine utilizes the Adapter Pattern to isolate AI platform-specific formatting requirements (Claude, OpenCode, Gemini, Roo, Kilo, Cursor, etc.) into standalone adapter modules, while Serializers handle raw data stringification (Markdown Frontmatter, YAML, JSON).

Architecture detail: docs/system-architecture.md.

Packages

Package	Purpose
@caesar/agents-core	Canonical schema, transpiler engine, emitters, and output validators
@caesar/cli	The caesar CLI: build, validate, and install
@caesar/{category}	Prebuilt agent artifacts per category, generated from dist/ at release

AGENTS.md fallback

For tools that read repository-level instructions instead of tool-native agent files, Caesar Harness Agent emits a shared AGENTS.md routing index. That fallback keeps role discovery consistent, but native targets remain the preferred path when a tool supports isolated agents, permissions, or model routing.

Contributing a new agent

Every agent goes through a 6-gate quality pipeline (see docs/g1-research-playbook.md and docs/system-architecture.md):

1. G1 Research — current primary sources plus corroborating engineering sources.
2. G2 Draft — copy agents/_template/canonical-agent-template.md; fill the frontmatter superset and all 6 body sections.
3. G3 Self-review — tool minimalism, model tier, crisp routing, no sibling-trigger overlap.
4. G4 Expert review — measurable rubric, read-only poka-yoke for auditors.
5. G5 Transpile + validate — caesar validate --strict must pass.
6. G6 Smoke test — emitted files load and parse for every target.

Local validation:

pnpm build
pnpm test
pnpm validate
pnpm lint

FAQ

What are expert coding subagents?

They are specialized agent definitions that tools can route to for focused work such as code review, infrastructure, data engineering, or security analysis.

Can these agents work natively in OpenCode or Codex?

Not directly as the same file. Caesar Harness Agent solves that by compiling one canonical source into native OpenCode, Codex, and other target formats.

What is AGENTS.md used for?

AGENTS.md is the shared repository instruction index for compatible tools and fallback routing. It is useful for consistency, but it does not replace native tool support where native agents are available.

How does Caesar Harness Agent prevent prompt drift?

The canonical agent is the only source that maintainers edit. Generated outputs are rebuilt and validated instead of hand-maintained per tool.

Is Caesar Harness Agent a prompt marketplace?

No. It is a harness, transpiler, and curated software-engineering agent library for maintainers who want repeatable multi-tool outputs.

Community

• Contributing guide
• Security policy
• Code of Conduct
• License

License

GPL-3.0-only. See LICENSE.