Pick a profile. Install the skills and agents your team needs. AgentShield scans every session. The GitHub App turns repo history into reusable defaults.
/plugin marketplace add https://github.com/affaan-m/ECC
/plugin install ecc
npm i -g ecc-universal
Prefer a local profile? Pick core, developer, security, or full in the install builder, or jump to GitHub App setup.
Watch ECC analyze a repository, extract workflow patterns, and turn them into reusable skills and defaults
Works across Claude Code, Codex, Cursor, and OpenCode
The repo drives distribution, AgentShield provides protection, and ECC 2.0 is the operator layer that sits above individual harnesses.
The repo stays the front door. Skills, commands, hooks, adapters, demos, and install profiles keep ECC useful before anyone pays.
Protection lives beside distribution, not behind it. Teams can audit configs locally, then add automation, policy packs, and enterprise reporting only where needed.
ECC 2.0 is shipped: a local-first control plane for observability, orchestration, session management, and cross-harness operations above the underlying tools.
One install path. Start free on public repos, move to Pro for private repositories and active seats, then scope Enterprise when rollout and governance become the real job.
Best for OSS maintainers and evaluation on public repos.
For teams that need private repo coverage and active-seat billing without changing the review flow.
For organizations that need procurement, governance, and rollout support on the same install path.
See how ECC upgrades the same workflow without forcing a tool reset or hiding the open-source layer.
The GitHub App is the conversion surface, but these are the workflows that make the ECC ecosystem sticky: battle-tested skills, agents, security tools, and automation helpers you can adopt in minutes.
Opus 4.6-powered security auditor for AI agent configs. Use it locally, in pull request review, or as the foundation for policy packs and enterprise reporting.
Test-driven development enforcer. Keep tests first, coverage visible, and implementation reviewable instead of relying on ad hoc prompts.
AgentAutomated review for quality, security, and maintainability. A strong default when teams want consistent standards across generated changes.
SkillOWASP Top 10 checklist, secret detection, and secure coding patterns for teams that want security review built into the same harness.
PluginCreate hooks conversationally. Describe the guardrail you want and generate the enforcement config without hand-authoring every rule.
SkillInstinct-based system that observes sessions and creates atomic behaviors with confidence scoring, so the harness improves with repeated work.
AgentFeature implementation planning that breaks down complex work into phases, dependencies, and execution order for both humans and agents.
Start with the GitHub App for the fastest repo-level setup, or use the OSS install builder for local harnesses.
Grant read access to git history and write access for PRs. Public repos can start free.
Type /ecc-tools analyze on an issue in your repo. The app audits the repo harness and turns repeatable patterns into guidance.
Merge the generated SKILL.md and instincts.yaml only after review. Compatible harnesses pick them up automatically.
Prefer local-only first? Jump to OSS install, pick a profile, and add the GitHub App later when repo automation matters.
Reviewable security for agent configs, hooks, MCP servers, and prompt surfaces before they become team defaults.
Scan CLAUDE.md, .cursorrules, agents.json, hooks, and MCP surfaces with the same open scanner you can later automate in pull request review. The value is clear scope, high-signal findings, and a credible path into policy packs and enterprise rollout.
Scans CLAUDE.md, .cursorrules, agents.json, and custom configs for security weaknesses and misconfigurations.
Red-team agent crafts adversarial prompts to test for injection vulnerabilities in your agent instructions.
Blue-team agent validates that safety boundaries, permission scopes, and tool restrictions are properly enforced.
Generates a structured security report with severity ratings, specific findings, and remediation steps.
Sourced from public GitHub discussions, not invented avatars. These are short quotes from the actual community using ECC in the wild.
This works very well. The model stays on track, checks its work, fixes the code without you having to ask it, so results come back much faster.
An incredible resource for understanding agentic workflows in Claude Code.
The token optimization and memory persistence sections address a real problem.
Thanks for backing the OSS harness ecosystem and helping keep ECC public, maintained, and improving in the open.
AI code review in every pull request, backing the open harness ecosystem and the public install surfaces teams start with first.
AI code review that understands your whole codebase, backing the open harness ecosystem alongside the ECC install surfaces.
Thank you for helping keep the public OSS layer healthy and useful.
ECC Tools is one layer in a broader OSS system. These are the reusable surfaces teams keep after the first install.
Curated workflow patterns for planning, coding, review, and shipping.
Specialized subagents for planning, review, testing, and execution-heavy work.
Pre and post tool automation for checks, formatting, and fewer repeated mistakes.
Captures repeated behavior and turns it into reusable patterns with confidence scoring.
Open scanner for configs, hooks, MCP, and agent surfaces with automation layered on top.
Product, install, and rollout updates across the GitHub App, OSS layer, AgentShield, and ECC 2.0.
Quick answers to common questions about ECC Tools.
Start with the GitHub App for the fastest repo-level path, keep the OSS install where you want local control, and move into Pro or Enterprise only when private repos, policy, or rollout pressure become the blocker.
Need the local OSS path? Open the selective install builder, copy the installer below, and layer the GitHub App in later. Already paid and wiped local setup? Open account recovery instead of starting checkout again.
Built by Affaan Mustafa · Open source under MIT license