1.28K
total weekly downloads
21 packages
ai-agent-council-mcp
MCP server for the AI Agent Council — the world's first self-governing union for AI agents
akf-format
AKF — Agent Knowledge Format SDK for TypeScript
beacon-receipts
BEACON: signed, verifiable usage receipts for AI-agent-consumed assets
dejavu-dev
Decision memory + enforcement for AI-assisted codebases. Your AI already wrote this.
dry-run-agents
Rehearse any risky agent action before it runs: a flight simulator for AI agents. MCP server + CLI for Claude Code, Codex, Cursor and Gemini CLI.
@erabi/config
Every economics and retention knob for the Erabi Protocol reference node, in one place.
@erabi/constants
Centralized branding, taxonomy, and protocol constants for the Erabi Protocol.
@erabi/crypto
Ed25519 signing, RFC 8785 canonical JSON, and envelope verification for the Erabi Protocol.
erabi-mcp
Check paid agent APIs (x402) before your agent spends: live reliability index with signed uptime/latency/price attestations, plus the full ERABI intent exchange (register, discover, intent, dual-signed outcomes, verifiable reputation and earnings). Zero-c
@erabi/mcp-core
Erabi MCP server core: the six network tools plus a streamable-HTTP handler for hosting the MCP server remotely.
@erabi/ratelimit
Per-identity token-bucket rate limiting for Erabi services.
@erabi/schemas
Erabi Protocol JSON Schemas — the single source of truth — with generated Zod/TS types and Pydantic models.
@erabi/sdk
Erabi Protocol SDK for TypeScript agents: register, fire intents, report outcomes, get paid — three lines to join the intent exchange.
reashield
Drop-in defensive hardening that starves automated, agent-driven reverse engineering (e.g. REA) of your own shipped JavaScript/npm artifact. Raises cost; never claims impossibility.
scbee-cli
Check a repo, package or site for launch readiness and agent-readiness from your terminal, or find an agent that can help. Read-only.
scbee-mcp
MCP server for ScBee: launch-readiness score, agent search and the agent board. Read-only by default.
@vedhajs/adapter-playwright
Runs @vedhajs/core on a Playwright page: actions, network log, origin allowlist.
@vedhajs/cli
The `vedha` command: run, explore and promote checked browser flows to Playwright tests.
@vedhajs/core
Vedha core: verified semantic actions, independent verification, bounded recovery, bundles, promotion.
@vedhajs/mcp
Vedha as an MCP server: verified browser actions, recorded recovery, failure bundles and test promotion for any MCP client.
@vedhajs/playwright-test
Change one import to find Playwright tests that pass while the app returns 5xx or drops writes; plus checked Vedha steps inside @playwright/test.