convez.ai
AgenteSegurançaAvançado
Jott2121

Agent Gate – Barreira de Qualidade para Agentes IA

Jott2121/agent-gate

Gate fail-closed + ledger hash-chained para disciplina em workflows de agentes IA

Ver no GitHub
InicialPluginScore · de 0 a 100

Um servidor MCP que implementa um portão de qualidade determinístico e fechado por padrão para agentes de IA. Fornece verificação fail-closed (prova explícita obrigatória), revisão independente refute-first e um ledger tamper-evident baseado em hash SHA-256 para rastreabilidade honesta de cada decisão do agente.

Instalação

$npx convez-ai Jott2121/agent-gate

1 comando no terminal: detecta o formato do repositório (marketplace ou plugin avulso) e instala via Claude Code. Requer Node 18+ e o Claude Code instalado.

Métricas

Estrelas
3
Forks
1
Acompanhando
0
Saúde
57%
Último commit
há 1 dia
Versão
v0.1.1
Issues abertas
1
Contribuidores
1

Casos de uso

  • Implementar portões de qualidade determinísticos em workflows de agentes IA antes de ações irreversíveis
  • Criar trilha de auditoria tamper-evident para decisões de agentes com integridade criptográfica
  • Enforçar aprovação humana mandatória em ações críticas de agentes antes de execução

Veredito

Ferramenta robusta de infraestrutura para governança de agentes IA, direcionada a arquitetos que precisam de garantias determinísticas e auditáveis. Implementação séria da filosofia Fleet Mode com estruturas criptográficas genuínas; exige integração técnica consciente.

README

A tradução para o português deste README ainda está na fila.

agent-gate: gate an AI agent's work before it ships with deterministic checks, refute-first review, tamper-evident receipts

agent-gate

ci License: MIT Python MCP

An MCP server that lets an AI agent gate its own work before it claims "done": deterministic checks, then an independent refute-first review, then a tamper-evident honest receipt.

Agents that grade their own homework ship low-quality output. agent-gate turns that discipline into tools an agent must actually pass: a fail-closed checklist and an append-only, hash-chained receipts ledger. It is Fleet Mode, my agent-orchestration doctrine, made into a runnable tool. Receipts over hype, enforced by the data structures.

agent: "done!"  ->  verify_gate(evidence)  ->  { passed: false, blocking: ["independent_refute_review", "no_secrets"] }

Why

The expensive failures in agent systems are the silent ones: a model update degrades output, a change quietly breaks a workflow, an agent declares success while the work is wrong. The fix is not a smarter model. It is a gate the agent cannot talk its way past:

  • Fail-closed. A check counts as satisfied only if it is explicitly true. Missing proof is not proof. (Mirrors a promotion gate, not an informal check.)
  • Tamper-evident receipts. Every decision is recorded as (decision, metric, value, verdict) linked into a sha256 chain. Edit or delete any past receipt and verify_chain() returns false. The honest log is enforced by the structure, not by good intentions.
  • Human-gated by default. "Any irreversible/outward act got human approval" is a required check. Agents draft, humans approve.

Tools (over MCP)

ToolWhat it does
gate_checklist(name="ship")Returns the checklist the agent must satisfy before claiming done.
verify_gate(evidence, name="ship")Evaluates evidence fail-closed and returns {passed, blocking}.
record_receipt(decision, metric, value, verdict)Appends an honest, hash-chained receipt; returns it.
read_receipts()Returns every receipt plus whether the chain is intact.

The default ship gate encodes Fleet Mode: deterministic_checks_pass, independent_refute_review, no_secrets, human_gated_if_irreversible, honest_receipt_logged.

Install & wire into an MCP client

pip install mcp-agent-gate   # or: pip install -e . (from source)

Add it to your MCP client (Claude Desktop / Claude Code) config:

{
  "mcpServers": {
    "agent-gate": { "command": "python", "args": ["-m", "agent_gate.server"] }
  }
}

Now your agent can call verify_gate(...) before it tells you it is finished, and you get a tamper-evident trail of what it decided. Receipts persist to ~/.agent-gate/receipts.jsonl (override with AGENT_GATE_LEDGER).

Use it directly (no MCP client needed)

from agent_gate.gate import DEFAULT_SHIP_GATE
from agent_gate.ledger import Ledger

res = DEFAULT_SHIP_GATE.evaluate({
    "deterministic_checks_pass": True,
    "independent_refute_review": True,
    "no_secrets": True,
    "human_gated_if_irreversible": True,
    # honest_receipt_logged missing  ->  fail-closed
})
print(res.passed, res.blocking)   # False ['honest_receipt_logged']

led = Ledger("receipts.jsonl")
led.append(decision="ship v0.1", metric="tests", value="pass", verdict="shipped")
print(led.verify_chain())         # True  (until someone edits the log)

Design

  • Tested, stdlib-only core. agent_gate/gate.py (fail-closed checklist) and agent_gate/ledger.py (hash-chained receipts) are pure stdlib: fast to read, fast to trust. agent_gate/server.py is a thin MCP adapter over them (the one runtime dependency: mcp).
  • Tests pass on Python 3.11-3.13 (see CI). The MCP tools are tested by calling them, not just importing.

Tests

pip install -e ".[dev]" && python -m pytest -q

Contributing

See CONTRIBUTING.md.

About

Built by Jeff Otterson (Jott2121). agent-gate operationalizes the gating discipline from bow (an autonomous all-Claude chief-of-staff agent) and the Fleet Mode doctrine. Siblings in the same line: rag-guard and agent-cost-attribution. MIT licensed.

Ver o original no GitHub

Bom para

Relacionados