npm

Search results

98 packages found

An open-source SDK for AI agent safety

published version 0.1.7-rc2, 3 months ago2 dependents licensed under $MIT
43,708

Deterministic AI code guardrail engine — 37 built-in rules, hybrid YAML/function DSL, MCP server, .cursorrules sync, sqlite cache.

published version 1.0.0-beta.2, a month ago1 dependents licensed under $MIT
2,622

Shared core for aicqtools — tree-sitter parser, sqlite incremental cache, SARIF/JSON/text reporters, config loader, i18n.

published version 1.0.0-beta.2, a month ago4 dependents licensed under $MIT
3,012

aicq command-line interface — `aicq check`, `aicq mcp`, `aicq sync-ai-rules`, `aicq provenance`, `aicq docs build`.

published version 1.0.0-beta.2, a month ago0 dependents licensed under $MIT
2,455

Runtime rule enforcer for AI coding agents. Reads CLAUDE.md / AGENTS.md / .cursorrules and enforces them via Claude Code and Cursor hooks, with a deterministic safety baseline.

published version 1.9.0, a month ago0 dependents licensed under $MIT
2,251

User rule authoring SDK for aicqtools — `defineRule()` helper plus Rule / RuleContext / Visitor types.

published version 1.0.0-beta.2, a month ago2 dependents licensed under $MIT
2,446

CLI and MCP servers for Cutline, including SlopBurn: a product quality engineering roguelike RPG for vibecoding workflows.

published version 0.38.6, a month ago2 dependents licensed under $MIT
1,066

Pharos Skill Engine-compatible transaction safety firewall for AI agents.

published version 1.5.0, 3 days ago0 dependents licensed under $MIT
1,802

Kill Switch for coding agents — stop runaway Claude Code / Cursor / Aider sessions from racking up an LLM bill. Native hook + token-metering proxy with per-session and daily-rolling budgets.

published version 0.1.11, 4 days ago1 dependents licensed under $MIT
1,352

AI Output Verification Toolkit — Local-first LLM safety, hallucination detection, PII redaction, prompt injection defense, and runtime monitoring. Zero telemetry. OWASP LLM Top 10 aligned.

published version 1.5.2, 4 months ago0 dependents licensed under $MIT
1,101

CLI and MCP servers for Cutline, including SlopBurn: a product quality engineering roguelike RPG for vibecoding workflows.

published version 0.15.1, a month ago3 dependents licensed under $MIT
926

MCP server that proxies all agent actions, classifies risk, enforces natural-language policies, and logs to a local SQLite audit DB. View activity in the hosted dashboard at roxy-agent-web.vercel.app.

published version 0.5.6, 10 days ago0 dependents licensed under $MIT
1,292

Entropy0 MCP server — source trust and URL safety tools for AI agents

published version 0.1.5, 25 days ago0 dependents licensed under $MIT
901

Core types, chain orchestration, and utilities for the Guardrail Chain framework

published version 0.1.0, a month ago2 dependents licensed under $MIT
572

CLI and MCP servers for Cutline, including SlopBurn: a product quality engineering roguelike RPG for vibecoding workflows.

published version 0.5.3, 15 days ago1 dependents licensed under $MIT
707

Observability interfaces (logging, metrics, tracing) for the Guardrail Chain framework

published version 0.1.0, a month ago2 dependents licensed under $MIT
578

Bindings-only Newton guardrail policy pack (zod schemas + deployments). No PolicyPack export yet — use with NewtonShield.guardedCall until pack.ts lands.

published version 2.0.7, 11 hours ago0 dependents licensed under $Apache-2.0
661

Built-in guardrail implementations for the Guardrail Chain framework

published version 0.1.0, a month ago0 dependents licensed under $MIT
402

Two-way guardrails for AI agents — gate inputs, outputs, and actions

published version 0.2.0, 24 days ago0 dependents licensed under $MIT
460

Guardrail plugin for OpenClaw that blocks prompt injection, jailbreak attempts, and sensitive credential exfiltration before agent dispatch.

published version 1.0.5, a month ago0 dependents licensed under $MIT
305