Provisions a LastID-issued agent credential on first session start. Approved by the operator in their LastID wallet; the issued SD-JWT VC + Ed25519 seed are persisted to the host keychain. Steady-state sessions skip provisioning.
Cryptographic identity for AI agents. The plugin provisions a
LastID-issued LastID.Agent.Base verifiable credential the first time
an agent runtime starts on a host — approved one-time in the operator's
LastID wallet — then runs authenticated on every subsequent session.
The plugin installs into:
codex mcp add)MCPServerStdioAll four paths bind the same underlying lastid-agent CLI + MCP server.
/plugin marketplace add GetTrustedApp/lastid-agent
/plugin install lastid-agent
On the next session start, the plugin reports provisioning status. If no agent credential exists yet, run:
/lastid-agent:provision
You'll be asked for your LastID canonical DID (did:lastid:z…). The
CLI then surfaces a verification URL + user code; open the URL on the
device that holds your LastID, approve the agent in your wallet, and
the issued SD-JWT VC is persisted to the host keychain.
npx -y @lastid/agent serve --http 0.0.0.0:8787
/mcp URL.(For local-only testing, expose the HTTP server via a tunnel — e.g. Cloudflare Tunnel or ngrok — and use the public URL.)
codex mcp add lastid-agent -- npx -y @lastid/agent serve
Then run lastid-agent provision once to register this host's agent
with your LastID wallet.
Python:
from agents.mcp import MCPServerStdio
server = MCPServerStdio(
name="lastid-agent",
params={"command": "npx", "args": ["-y", "@lastid/agent", "serve"]},
)
JS/TS:
import { MCPServerStdio } from '@openai/agents';
const server = new MCPServerStdio({
name: 'lastid-agent',
command: 'npx',
args: ['-y', '@lastid/agent', 'serve'],
});
lastid-agent provision # one-time: provision this host's agent identity
lastid-agent status [--json] # report provisioning state
lastid-agent show # print the stored agent VC (debug)
lastid-agent serve # MCP server on stdio (Claude/Codex/Agents)
lastid-agent serve --http :8787 # MCP server on HTTP (ChatGPT)
Flags accepted by provision:
--parent-human-did did:lastid:z… REQUIRED (or env LASTID_PARENT_HUMAN_DID)
--idp <url> Default: https://human.lastid.co
--runtime <name> Default: lastid-agent-cli
--project-hint <hex> Optional SHA-256 prefix
--scope <slug> Default: main
--force Overwrite existing keychain entry
The first release surfaces a single tool so the install + handshake work end-to-end across runtimes:
lastid_whoami — returns the agent's DID, parent human DID,
capabilities, and expiry. Reports { provisioned: false } when no
agent credential is present.Memory (lastid_memory_*), signing (lastid_sign), and messaging
(lastid_message_*) tools land in subsequent releases on the same MCP
server.
.claude-plugin/marketplace.json # Claude Code marketplace index
plugins/lastid-agent/
.claude-plugin/
plugin.json # plugin manifest
hooks.json # SessionStart wiring
.mcp.json # Claude Code MCP autodiscovery
bin/lastid-agent.js # CLI: provision / status / show / serve
hooks/session-start.js # thin CLI shell (no native keychain calls)
commands/provision.md # /lastid-agent:provision slash command
lib/
agent-provisioning.js # OID4VCI client
keychain.js # OS-keychain adapter
mcp-server.js # MCP server, stdio + HTTP
tests/
package.json # bin: lastid-agent
Apache-2.0. © LastID.
Admin access level
Server config contains admin-level keywords
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
npx claudepluginhub gettrustedapp/lastid-agent --plugin lastid-agentGate destructive AI operations behind LastID verifiable credential presentation. Detects dangerous tool calls and requires DID-based identity verification before proceeding.
Upstash Context7 MCP server for up-to-date documentation lookup. Pull version-specific documentation and code examples directly from source repositories into your LLM context.
Production-grade vibe coding suite. Structured workflows from discovery to deployment: start, build, ship, fix, refactor.
20 SEO/GEO skills and 5 commands on one shared contract for keyword research, content creation, technical audits, schema markup, monitoring, quality gates, entity truth, and campaign memory.
Institutional-grade CRE skills library covering the full investment lifecycle: deal screening, underwriting, structuring, due diligence, capital markets, market research, asset management, leasing, investor relations, development, disposition, sourcing, tax, ESG, portfolio strategy, and daily operations. Includes Monte Carlo simulation, SEC Reg D compliance, property management ops, deal attribution tracking, 1031 pipeline management, distribution notice generation, fund-raise LP negotiation tracking, and emerging manager evaluation. Ships expert subagents, Python calculators, workflow chains, and orchestrator pipelines for multi-agent acquisition workflows.
Unified capability management center for Skills, Agents, and Commands.