meta-harness

name	meta_harness
description	Meta-Harness orchestre les skills Botte comme des agents interchangeables dans un pipeline gouverné. Planifie → exécute en sandbox → review croisé → applique avec garde-fous. Inspiré d'Omnigent mais 100% Botte-native.
tags	["orchestration","pipeline","governance","sandbox","multi-agent"]

meta_harness — orchestration multi-agent

Un meta-harness qui orchestre les skills Botte (audit, fix, counter-audit, optimize, test) dans un pipeline gouverné. Chaque étape tourne dans son propre sandbox (subprocess, workdir isolé). Governance = approval gates + budgets + rollback.

Concept

USER: "audit + fix mon projet"
    │
    ▼
Meta-Harness
    ├── Orchestrator → Plan: [audit] → [review] → [fix] → [test]
    │
    ├── Sandbox 1: Porthos audit (skills/directives_audit)
    ├── Sandbox 2: Rochefort counter-audit (skills/cardinal)
    ├── Sandbox 3: d'Artagnan fix (skills/fix)
    └── Sandbox 4: run tests
    │
    ├── Governance → approval gate avant apply
    │
    └── Report → synthèse multi-agent

Agents disponibles

Agent	Skill Botte	Rôle
porthos	directives_audit	Audit initial
rochefort	cardinal	Contre-audit (red team)
d'artagnan	fix	Correction automatique
aramis	optimize	Optimisation token
conductor	conductor	Planification
security	security_scanner	Scan sécurité
fast_context	fast_context	Exploration repo

Usage

# Pipeline complet: audit → counter-audit → fix → test
python -m skills.meta_harness.cli run . audit counter-audit fix test

# Voir les pipelines disponibles
python -m skills.meta_harness.cli plans

# Lancement avec approval gate
python -m skills.meta_harness.cli run . audit fix --approval

# Voir l'état d'un pipeline
python -m skills.meta_harness.cli status <session_id>

# Rollback
python -m skills.meta_harness.cli rollback <session_id>

API Python

from skills.meta_harness import MetaHarness, PipelinePlan, Step, Sandbox

# Créer un harness
h = MetaHarness(workdir="/path/to/project")

# Planifier un pipeline
plan = h.plan(["audit", "counter-audit", "fix", "test"])

# Exécuter
session = h.execute(plan, approval=False)

# Voir le rapport
print(session.report())

Architecture

meta_harness/
├── orchestrator.py   — Planification pipeline + dispatch
├── runner.py         — Sandbox subprocess + workdir isolé
├── governance.py     — Garde-fous, budgets, rollback
├── session.py        — Persistance + historique
├── cli.py            — argparse CLI
└── test_meta_harness.py