wshobson/agents - Agent Plugin Marketplace¶

Source: wshobson/agents (9.1K+ installs for top skill) Category: Agent Orchestration / Plugins Harnesses: Claude Code, Codex CLI, Cursor, OpenCode, Gemini CLI, GitHub Copilot

Massive agent plugin marketplace with 94 plugins, 203 agents, 175 skills, and 109 commands. One source-of-truth (plugins/) compiled into idiomatic artifacts for six different agent harnesses. Each harness gets native artifacts - not lowest-common-denominator translations.

Installation¶

For Hermes (Claude Code-based)¶

# Add the marketplace
/plugin marketplace add wshobson/agents

# Install individual plugins
/plugin install python-development
/plugin install security-audit
/plugin install memory-forensics

For Other Harnesses¶

# Codex CLI
npx codex-marketplace add wshobson/agents

# Cursor
# Add marketplace, then /plugin install <name>

# Gemini CLI
gh repo clone wshobson/agents ~/agents && cd ~/agents
make generate HARNESS=gemini && gemini extensions install .

# OpenCode
gh repo clone wshobson/agents ~/agents && cd ~/agents
make install-opencode

Verify:

/plugin list | grep wshobson

Plugin Categories (94 Total)¶

Category	Count	Top Plugins
Development	25+	python-development, typescript-development, rust-development, go-development
Security	12	security-audit, memory-forensics, vulnerability-scan, threat-modeling
Infrastructure	15	terraform-iac, kubernetes-management, docker-orchestration, aws-architecture
Data & ML	10	sql-optimization, data-pipeline-design, ml-model-deployment
Architecture	8	system-design, api-design, microservices-patterns
Documentation	6	api-docs-generator, readme-writer, changelog-automation
Business & SEO	8	seo-audit, growth-strategy, market-research
Memory & Context	5	memory-safety-patterns, memory-forensics, context-optimization
Orchestration	16	multi-agent-coordination, full-stack-workflow, incident-response

Key Plugins for Hermes/CorpusIQ¶

Memory & Security Plugins¶

# Memory safety patterns for agent context management
/plugin install memory-safety-patterns

# Analyze agent memory for security issues
/plugin install memory-forensics

# Extract security requirements from specifications
/plugin install security-requirement-extraction

Orchestration Plugins¶

# Multi-agent coordination for complex workflows
/plugin install multi-agent-orchestrator

# Full-stack development with agent handoff
/plugin install full-stack-workflow

# Incident response automation
/plugin install incident-response

Development Plugins¶

# Python development with testing, linting, type checking
/plugin install python-development

# TypeScript/Node development
/plugin install typescript-development

# Infrastructure as Code
/plugin install terraform-iac

Architecture¶

plugins/
├── python-development/     # 90+ local plugins
├── security-audit/
├── memory-forensics/
├── ... (94 total)
├── agents/                  # 203 domain-expert agents
├── skills/                  # 175 modular knowledge packages
└── commands/                # 109 slash commands

Each plugin is isolated and composable. Agents, commands, and skills are auto-discovered. Progressive disclosure ensures skills only load when activated.

Harness Compilation¶

Source (plugins/) ──▶ Harness Generator ──▶ Native artifacts
                         │
                         ├── Claude Code (.claude-plugin/)
                         ├── Codex CLI (codex-marketplace format)
                         ├── Cursor (.cursor-plugin/)
                         ├── OpenCode (generated + symlinked)
                         ├── Gemini CLI (generated + transformed)
                         └── GitHub Copilot (copilot plugin format)

Common Workflows¶

Security Audit Workflow¶

# 1. Install security plugins
/plugin install security-audit
/plugin install vulnerability-scan
/plugin install memory-forensics

# 2. Run security audit
/security:audit --target ./src --output audit-report.md

# 3. Run memory forensics on agent context
/memory:forensics --session-id current --output memory-report.json

Multi-Agent Development¶

# 1. Install orchestrator + dev plugins
/plugin install multi-agent-orchestrator
/plugin install python-development
/plugin install typescript-development

# 2. Start coordinated workflow
/orchestrator:full-stack --spec ./spec.md --output ./output/

Troubleshooting¶

Symptom	Cause	Fix
`/plugin marketplace add` fails	Network/auth issue	Verify `gh` CLI is authenticated: `gh auth status`
Plugin not found	Marketplace not registered	Re-run `marketplace add` command
Skills don't activate	Plugin not installed	Check with `/plugin list`
Harness-specific failures	Generated artifacts stale	Re-clone and regenerate for your harness
Gemini CLI setup fails	Missing `make`	`sudo apt-get install make`

Comparison: Other Agent Marketplaces¶

Marketplace	Plugins	Harnesses	Hermes Support
wshobson/agents	94	6	Via Claude Code harness
anthropics/knowledge-work-plugins	11	2 (Cowork + Code)	Via Claude Code only
github/awesome-copilot	14 skills	1 (Copilot)	MCP generators work anywhere
vercel-labs/agent-skills	6	1 (Vercel)	Framework-specific

Recommendation: wshobson/agents has the broadest coverage. Use as the primary agent plugin source. Supplement with github/awesome-copilot for GitHub/MCP automation and anthropics/knowledge-work-plugins for enterprise workflows.

Roster Additions (Sep 29, 2026 sweep)¶

Six skill listings from the wshobson/agents marketplace surfaced in the Sep 29 sweep:

Skill	Installs	What It Does
`code-review-excellence`	29,048	Code review excellence standards
`api-design-principles`	29,000	API design principles and patterns
`python-design-patterns`	21,414	Python design patterns reference
`prometheus-configuration`	10,485	Prometheus monitoring configuration
`team-composition-patterns`	8,942	Team composition patterns for agent work
`hermes-tweet`	1,940	Hermes-native tweet composition and posting

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.

See what CorpusIQ connects