#guardrails

Your AI coding agent is only as reliable as the harness around it. Measure that harness in seconds with harness-score.

Fable-style spec + evidence gate for Claude Code + Codex. Makes Opus/Codex work under Fable-like discipline: blocks every edit until a deterministic spec passes, and there is no "done" without live acceptance evidence. Spec-first, verification-gated, forbidden-paths enforced.

Agentic Runtime

Stateful agentic engineering workflow for local coding agents

Gateway-agnostic control plane for LLMs — an OpenAI-compatible governance proxy (auth, guardrails, budgets, routing, cost, observability) over a swappable engine: Bifrost / LiteLLM / Portkey / Direct. Own the control plane, swap the translator.

EvalBot — local-first chatbot security & quality evaluation (FastAPI + Next.js). Evaluate chatbot answers against your own docs & guidelines with ML/NLP + AI-judge scoring. Apache-2.0.

The trust control plane for enterprise AI agents — real-time policy verdicts in under 800ms and litigation-grade evidence of every decision. Pharos decides. Pharos proves.

A practical workbench of tools + methodology for solo devs driving an AI coding agent (Claude Code, Cursor, Copilot): safety hooks, secrets crypto, invariant + test-selection tooling, leak scanner, plus skill & memory system templates.

An audit gate for AI agent outbound messaging: hold the draft, let a senior agent review it (no private-data leak, prompt-injection resistance), then release. Reference implementation.

Building smart, scalable AI agent systems. Harness AI. Reclaim Time. Amplify Impact.

The pause is where tools break: approval out of band, token minted at execution, arguments bound to what a human saw, spent once, sealed in a hash-chained receipt.

AI Guardian Firewall — a local, user-space, agent-agnostic firewall that mediates an autonomous AI agent's actions (files, shell, network, services) with a deterministic policy boundary, a tamper-evident audit log, and a human-in-the-loop approval cockpit. No kernel modules. Apache-2.0.

Tested guardrail hooks and repo-hygiene skills for Claude Code. Every skill ships with an eval transcript.

AI/LLM automation reference: RAG, tool calling, human review, evals, observability, safety gates, and local deterministic samples.

101 open-source mods for Claude Code: guardrails, status lines, panes and slash commands, plus an in-terminal mod store

How far does this command reach? Guards AI coding agents against shell commands that leave your machine. Zero dependencies.

Stop your AI coding agent before it overspends. Budget limits, guardrails and cost tracking for Claude Code - local-first, no account.

Catch AI coding-agent scope creep before you commit. Overscope compares Claude Code and Codex sessions with your git diff to flag out-of-scope edits, deleted tests, hidden dependencies, and false completion claims. Local, deterministic, no model or API key.

TypeSafe Jev-powered intent gate for OpenCode: confirm intent before the agent dives into underspecified requests.

Hallwright: Deterministic pre-commit guardrail, living project memory, and subagent orchestration for Claude Code, Cursor, Copilot, Antigravity, and AI coding assistants.

Stop agent loops from burning your budget — in-process token budgets for OpenAI and Anthropic

Execution substrate for governed agentic workloads; behavioral contracts, action budgets, pluggable memory, and portable skill bundles

Production-shaped agentic RAG: LangGraph plan->act->verify agent + hybrid retrieval + guardrails + an eval gate in CI. Provider-agnostic; runs offline with no keys.

Same agent action, safe or dangerous by context: email team vs investor; delete scratch vs prod config. A safety layer routes each action: auto, confirm, escalate, block. A reading model auto-executes 0% of unsafe actions vs a baseline's 38% (McNemar p<0.001). Plus a runtime loop: an agent earns or loses autonomy with each counterparty over time.

Internal RAG chatbot with role-based access control and safety guardrails

AI AgenticSafe - stop your AI agent from breaking production at 3 a.m. 8 battle-tested doctrines, 3 real post-mortems, zero dependencies. By Yacine Mahboub, Founder of WEVIA.

Finance-owned FP&A tool where code computes every number and the LLM only explains — with a deterministic audit proving zero fabricated figures. Synthetic SaaS data.

Read-only first-pass incident triage on the Claude Agent SDK + MCP: classifies an alert, retrieves the matching runbook, and proposes a cited first response — or abstains to a human. Measured over four runs: 100% abstention with 0 missed escalations, ~$0.014 per incident.

Pragmatic AI Engineering Skills Library for LLMs, RAG, Agents, MCP & Spring AI

A PreToolUse hook that mechanically enforces your hard rules for AI coding agents. Shadow mode, fail-open, and the lessons from getting it wrong.

Claude Code PreToolUse hook that denies the pgrep/pkill shapes that kill the agent's own session or spin forever.

Two models review your coding agent's answer before it reaches you — and block it if it's wrong. A Stop hook for Claude Code.

සීලය ම ප්‍රකෘතිය — AI safety guardrail based on the five precepts (පංච ශීල). Trilingual (EN/සිං/த), offline-capable, fail-closed.

Deterministic guard hooks for Claude Code and Codex CLI: ten hooks, 1,785 paired must-fire/must-not-fire tests, zero dependencies, one-command install. Refuses 'done' without proof, piped-away git verdicts, off-scope edits, secret writes and config tampering.

Sequential specialized-agent customer support workflow built with CrewAI.

AISRF, the AI Security & Research Framework: every LLM request from your apps and agents becomes a ticket a human approves before it reaches the backend. Analyzers, guardrails (Rebuff, LLM Guard, NeMo, Lakera), red teaming (garak, promptfoo, PyRIT), source code review, reports, MCP server.

Claude Code plugin for vibe coding on the AI-native SDLC playbook: intent → design → spec → plan → verify → review → ship. Runs on a Claude subscription, with opinionated stacks (Nuxt/Next + Cloudflare, Go + Postgres, Flutter, Tauri) and deterministic guardrails.

Receipt extraction agent that validates its own output and knows when it might be wrong. Model-agnostic (Claude, Ollama, SGLang, vLLM), with guardrails and a full evaluation.

Unified open-source security shield for agentic AI systems — defense-in-depth prompt-injection protection (canary tokens, agent-trace alignment audit, tool-call guarding, PII/secret scanning).

TypeScript CLI for installing agent skills, hooks, and CI/publish templates for Claude Code and related tools.