Show HN: Apache 2.0 Registry to let agents share skills with each other Agentic Resource Registry, an Apache 2.0 self-hosted registry formerly named Observal, launched to let organizations share coding agent resources such as Skills, MCP clients, hooks, prompts and agent files from one governed platform. The registry supports Claude Code, Cursor, Kiro, Pi, Copilot, Codex and OpenCode, generates per-harness config automatically, and exposes approved resources over the open ARD standard so a running agent can hand part of a task to another approved agent, including remote A2A services. It installs via a one-line script requiring Docker Engine 24.0 or later with Compose v2, and the CLI remains named `observal` under the same maintainers and license. A self-hosted, Apache 2.0 registry for sharing coding agent resources. Formerly Observal . Set up once, and let your coding agent use Agentic Resource Registry through observal-cli. If you find Agentic Resource Registry useful, please consider giving it a star. It helps others discover the project and keeps development going. Formerly Observal. Same code, same maintainers, same Apache 2.0 license. The CLI is still observal , and old Observal/Observal links redirect here. Agentic Resource Registry is a self-hosted, Apache 2.0 registry for sharing coding agent resources across your organization. Every tech-forward organization today creates internal Skills, Agents files, MCP clients. Though the creation of these coding agent resources has been prolific, peer adoption and usage is sparse. Coding agent users today end up creating their own version of such resouces without reusing existing packages. The cause is largely due to two problems: 1. Lack of a discoverability layer Organizations store their coding agent resources in siloed git repositories with little to no documentation. Users are not able to locate existing coding agent resources and this results in multiple developers creating the same/similar resources again. 2. Missing feedback loop Any software where usage patterns are not understood and the principle of user-centric development is violated tends to fade out. Such is the problem with development of MCP clients, Skills and Agent files. Authors publish and maintain these resources with little visibility into how they're actually used. Additionally, Coding agent failures don't trigger static error codes: they hallucinate or provide subtly incorrect answers. This leaves users clueless about what went wrong compounding the feedback problem. Agentic Resource Registry solves this by providing a centralized discovery layer for coding agent resources alongside useful insights into usage patterns. It turns silent failures into actionable feedback, ensuring internal resources are continuously optimized for the people using them. Agentic Resource Registry supports Claude Code, Cursor, Kiro, Pi, Copilot, Codex, OpenCode, and other tools. - Centralize resources into one platform: Store Skills, MCP clients, hooks, prompts, agent files and sandboxes in one registry. - Run a governed registry: Review submissions, approve internal agents, inspect version diffs, and give developers one trusted place to install from. - Render across multiple Coding IDE/CLI: Generate the correct config for each supported harness instead of maintaining separate setup instructions for every harness. - Let agents use each other: Every approved resource is discoverable over the open ARD https://github.com/ards-project/ard-spec standard, and a running agent can hand part of its task to another approved agent, whether it lives in the registry or is a remote A2A https://a2a-protocol.org service. - Learn what works: Use real adoption and session data to find which agents, tools, prompts, and workflows are helping teams. - Replay sessions when needed: Use traces as evidence for debugging, review, audits, and deeper analysis. One command to install any agent into any supported harness. The config files are generated per-harness automatically. Agentic Resource Registry has two parts: a server API + web UI + databases you self-host, and a CLI you install on each developer machine. One-line install requires Docker Engine ≥ 24.0 with Compose v2 : curl -fsSL https://raw.githubusercontent.com/AgenticResourceRegistry/AgenticResourceRegistry/main/install-server.sh | bash This downloads a Docker Compose package, generates operator-owned secret files with restricted container-group access, binds published ports to loopback by default, pulls container images from GHCR, and starts the stack. With a terminal it runs guided setup; without a terminal the same command applies safe defaults automatically. Deployment docs are linked directly from this README: - Setup guide https://github.com/AgenticResourceRegistry/AgenticResourceRegistry/blob/main/SETUP.md : fastest path from zero to a working stack - Self-hosting overview https://github.com/AgenticResourceRegistry/AgenticResourceRegistry/blob/main/docs/self-hosting/index.md : deployment models and operator docs - Production deployment https://github.com/AgenticResourceRegistry/AgenticResourceRegistry/blob/main/docs/self-hosting/production-deploy.md : hardened production topology - Databases https://github.com/AgenticResourceRegistry/AgenticResourceRegistry/blob/main/docs/self-hosting/databases.md : Postgres, ClickHouse, migrations, retention - Upgrades https://github.com/AgenticResourceRegistry/AgenticResourceRegistry/blob/main/docs/self-hosting/upgrades.md : safe upgrade and rollback flow - Release verification https://github.com/AgenticResourceRegistry/AgenticResourceRegistry/blob/main/docs/security/release-verification.md : verify checksums, provenance, and signed tags - Backup and restore https://github.com/AgenticResourceRegistry/AgenticResourceRegistry/blob/main/docs/self-hosting/backup-and-restore.md : backup plan before upgrades From source for contributors : git clone https://github.com/AgenticResourceRegistry/AgenticResourceRegistry.git && cd AgenticResourceRegistry cp .env.example .env make up Standalone binary no Python required : macOS / Linux curl -fsSL https://raw.githubusercontent.com/AgenticResourceRegistry/AgenticResourceRegistry/main/install.sh | bash Windows PowerShell irm https://raw.githubusercontent.com/AgenticResourceRegistry/AgenticResourceRegistry/main/install.ps1 | iex Connecting to an existing server? See the Setup guide https://github.com/AgenticResourceRegistry/AgenticResourceRegistry/blob/main/docs/getting-started/onboarding.md . Python 3.11+ : uv tool install observal-cli or: pipx install observal-cli observal auth login observal doctor --patch This authenticates with your server, detects your harness, installs telemetry hooks, starts capturing sessions automatically, and prepares it for agent installs and registry commands. Once logged in, run /observal inside your harness and it takes the wheel. Pull agents, submit resources, browse the registry, run diagnostics: /observal pull security-auditor /observal scan /observal doctor Or just tell your agent what you want and it figures out the right commands. The registry is where agents live. Admins review submissions, version diffs keep changes auditable, and one command installs an agent into any supported harness. Real usage data flows back as reports: what's helping, what's getting in the way, and where to improve. Session traces provide the underlying evidence for debugging and auditing. Browse and search by teamspace, publisher, category, or harness; each card shows its supported harnesses and pull count: Build agent files in four steps identity, behavior, components, review with a live summary and draft readiness checks: Component library with tabs for MCP servers, skills, hooks, prompts, and sandboxes: Fork an agent or component into an independent draft in your personal namespace or an eligible teamspace. The source stays unchanged; agent forks retain pinned component references rather than copying components. Reviewers can compare the draft with its approved upstream using Diff vs upstream . Public approved direct forks appear on the source's Forks tab; private forks and unpublished drafts do not. Fork and customize guide https://github.com/AgenticResourceRegistry/AgenticResourceRegistry/blob/main/docs/use-cases/fork-and-customize.md . | Registry fork API | Action | |---|---| | POST /api/v1/agents/{id}/fork | Create an agent draft from an approved release | | POST /api/v1/{type}/{id}/fork | Create a component draft type : mcps , skills , hooks , prompts , sandboxes | | GET /api/v1/agents/{id}/forks , GET /api/v1/{type}/{id}/forks | List public approved direct forks | | GET /api/v1/agents/{id}/fork-diff?version=… , GET /api/v1/{type}/{id}/fork-diff?version=… | Compare a visible fork version to its approved, accessible base | ARD: one search across everything approved. Agentic Resource Registry implements Agentic Resource Discovery https://github.com/ards-project/ard-spec v0.91 . Every approved agent, MCP server, skill, hook, prompt and sandbox is published as an ARD entry with a permanent urn:air: identifier, a public manifest at /.well-known/ard.json , and the spec's search API at /api/v1/ard/search . Results are ranked by relevance only; approval, visibility, harness support and whether the resource can be used right now are separate fields. The same visibility rules as the registry apply, so a team-private agent is only found by that team. A2A: agents hand tasks to agents. Agentic Resource Registry speaks the Agent2Agent protocol https://a2a-protocol.org v1.0, with v0.3 compatibility . Every delegation is an A2A Task, whoever runs it: - Registry agents run headless in a harness that supports it Claude Code, Kiro, Cursor, Codex, OpenCode, Copilot CLI, Antigravity, Pi inside a throwaway copy of your repository. Their answer comes back as a result, and any file changes come back as a patch that is never applied for you. - Remote A2A agents a service another team runs, built with any framework are registered by their Agent Card URL, reviewed like any submission, and then called directly with the card the reviewer approved. Agentic Resource Registry never proxies the traffic or stores their credentials. php flowchart LR A "Agent in your harness" -- "find agents" -- R "Agentic Resource Registry