#audit

Deterministic authorization + tamper-evident audit for AI agents, humans, and workloads — one principal type, safety invariants machine-checked by an SMT solver (cvc5), every decision independently verifiable offline. Pure Rust.

Open-source loop engineering for auditable business briefings: claims, evidence, gates, findings, repairs, and human review.

CLI/MCP agent surface for Enji Guard

Agent-safe action and workflow runner for TypeScript backends.

Audit long-form LLM output for factual accuracy, hallucination risk, source quality, and actionable edit suggestions.

A portable AI-agent skill for finding project blind spots, unknown unknowns, hidden risks, and missing decisions, with a durable ledger for repeat audits across Claude Code, Claude desktop/Cowork, Codex, and OpenCode.

Universal security audit skill for Claude Code — OWASP Top 10 (2021) + Russian security standards (ГОСТ Р 56939, ФСТЭК, ФЗ-152) + STRIDE threat modeling. Multi-stack: Node/Python/Go/Rust.

Skills Audit — 技能审查·画像·评分·优化 / Universal agent tool auditor — profile, score, optimize

A living spec defending against the 26 categories of mistakes LLMs systematically make in Rust. Claude Code skill + 3 slash commands (rust-audit, rust-fix, rust-plan) built on a single source of truth.

Read-only Rust control plane for ocserv fleets, with iroh EndpointID trust, a SQLite-backed CLI controller, and audited node agents.

Authenticity, integrity, and capability audit for AI API gateways — verify the model behind any Anthropic / OpenAI / Google compatible endpoint.

See your store the way AI shopping agents do: scored, product-by-product agent-readiness audits for any e-commerce catalogue

A practical map for running a legal digital business in the EU: 206 sourced compliance checks, audit method, and AI-agent-ready requirements (GDPR, ePrivacy, AI Act, DSA, DMA)

Hermes-inspired security control plane for LLM agents over private DMs: allowlist, PIN, kill switch, rate limits, injection heuristics, secret redaction, audit log.

The Trust Stack for Agent Apps — independently installable Python libraries that make AI agents reliable, auditable, and accountable. Because agent failures are trust failures.

Switchback Governance · 折返治理 — the human-in-the-loop governance layer for multi-agent teams (GOAI 2026 Track 1)

NOA — the Agent Action Receipt. Open governance layer for AI agents: gate risky actions before they run, emit a tamper-evident receipt anyone can verify offline. Early access · Apache-2.0.

Audit your A2A agent card across L1-L4 trust dimensions on every PR. GitHub Action wrapping @agentlair/a2a-trust-audit.