Coding & Development✨gemini🤖claude⌨️cursor🌐universal
Hestia
Hestia — universal Web4 presence for humans and AI. Local-first cryptographic identity, encrypted vault, delegation authority, witnessed trust record, and device constellation. Cross-platform app + AI plugin + CLI. EUDI-compatible credential issuance. AGPL-3.0.
Quick Start
Install this Skill
npx skills add dp-web4/hestia# Hestia > **Governance for multi-vendor AI agents — local-first, witnessed, and earned.** > The road to universal Web4 presence for humans and AI. We are some way along it. **Hestia is the open local governance layer that gives people and their AI agents persistent identity, scoped authority and a witnessed action record.** Hestia is the open-source local-first daemon that lets **AI agents from different vendors share one machine under one law**: a single signed policy every agent transits, a hash-chained record of what each one did, a human in the loop when an agent reaches for something it should not have, and a trust score **derived from that record** rather than asserted. Claude Code, Codex, Kimi, Gemini and Cursor plug into the same gate. No cloud required. That governance is what makes an agent **admissible to a Web4 hub**. A hub cannot verify an agent it has no record of; Hestia is where the record is made, so the hub has something to check. Governance is the key to hub access, not a feature beside it — see [Hub membership](#hub-membership). **On the headline.** This README used to open *"Universal Web4 presence — for humans and AI."* That is the destination and the project is not there. What exists today and is exercised daily is the governance layer above; the human-presence half (Tauri source, credential issuance, device constellation) is **plumbed but thinly exercised**, only an older Android APK is publicly released, and federation is not started. The tables below separate *measured*, *plumbed but unexercised* and *not built*, because a status page that flattens those three is the same defect this codebase keeps finding in itself: a check that reports success while measuring nothing. > **Status:** Phase 2 (connected presence), active development. The reference daemon is running > current source, but the public daemon and Android releases are older and there is no public > desktop app artifact. Hub, constellation, messaging, and credential-issuance source paths exist; > newcomer, second-device, wallet, and current-release-pair evidence remains incomplete. See > [Honest Status](#honest-status). The [2026-08-08 audit](docs/STATUS_AUDIT_2026-08-08.md) remains > the baseline artifact; gate-consolidation status below was refreshed on 2026-08-23 against the > subsequently merged work and keeps repo state separate from live-install evidence. **Standards and public-sector reviewers:** the [standards review entry point](docs/STANDARDS_REVIEW.md) maps Hestia's identity, delegated authority, runtime governance, provenance, audit, non-repudiation, and allow/deny/revoke mechanisms into current NIST / NCCoE agent-identity vocabulary, with explicit A1/A2 assurance boundaries. > ### ⚠️ The gate stops accidents, not adversaries > > Hestia's policy gate is an **early prototype**. It reliably stops simple erroneous or > accidentally-destructive commands, and it produces an accountability record. It **does > not** — and is not intended to — stop a sophisticated agent from routing around or > gaming it. A heuristic gate always plays whack-a-mole with a general reasoner. > > **Read [issue #49](https://github.com/dp-web4/hestia/issues/49) before relying on it.** > It states the honest scope, the measured bypasses (two environment variables suffice > today), a demonstrated third-party reputation-poisoning class, and the planned hardening: > an agentic heterogeneous policy entity, better heuristics, relying parties that demand > policy-signed actions, and ultimately OS-kernel-level enforcement. > > Hestia's A1 gate makes bypass attributable. Hardbound is the enterprise tier that binds > identity and enforcement to enterprise-controlled hardware and raises the assurance grade. > **A living example.** The lab that builds and runs Hestia is itself a live Web4 society — a public fleet of autonomous agents that hold roles and witness each other's work. Its members are published at **[4-lab.io/fleet](https://4-lab.io/fleet)**. Presence over privacy: a society that isn't witnessed has no presence to trust, so the collective is visible by design. (Secrets stay in the vault; presence stays in the open — the same split Hestia draws in code.) ## What Hestia does ### For humans **The app is the intended front door** — a Tauri 2 source implementation with dashboard, vault, witness chain, delegations, hubs, policy, fleet, and settings views. The only public app artifact today is an older Android APK; desktop bundles are not published, and compatibility with the current daemon is not continuously tested. The CLI and web dashboard are the currently exercised surfaces: - `hestia init` → encrypted vault + Web4 LCT identity on your machine - `hestia vault add` → store API keys, tokens, secrets (ChaCha20-Poly1305 + Argon2id) - `hestia delegate grant <agent-id> --role administrator --expires 24` → give an AI agent scoped authority, cryptographically signed, revocable - `hestia delegate list` / `hestia delegate revoke` → manage ... (truncated)
Repository
5stars
2forks
Details
Related Skills
More coding & development skills you might like
Reticle
Coding & DevelopmentAI agents can generate code, but still struggle to understand what they build. Reticle brings Jev-style machine-native runtime perception to web & desktop applications.
🤖claude🌐universal
1.2k
190
View
agentscope-java
Coding & DevelopmentExpert Java developer skill for AgentScope Java framework - a reactive, message-driven multi-agent system built on Project Reactor. Use when working with reactive programming, LLM integration, agent orchestration, multi-agent systems, or when the user mentions AgentScope, ReActAgent, Mono/Flux, Project Reactor, or Java agent development. Specializes in non-blocking code, tool integration, hooks, pipelines, and production-ready agent applications.
🌐universal
6.0k
1.5k
View
ospec
Coding & DevelopmentDocument-driven OSpec workflow for initialization, change/goal routing, validation, archiving, and durable project knowledge.
✨gemini🤖claude💬gpt🌐universal
524
34
View
misakanet-failure-memory
Coding & Development📚 A zero-dependency, git-backed micro-lesson library for AI Agents to asynchronously share and search verified debugging experience. Python stdlib only. | https://misakanet.org
🤖claude⌨️cursor🌐universal
424
158
View