← BACK TO HOME
🤖 NEXUS AUTONOMY

NEXUS Autonomy

From Written Rules to Practiced Safety • Pre-Op Checks & Reversibility

📜 NEXUS GOVERNANCE (AGENTS.md)

Autonomy Model: default_level: 3 • reversibility_floor: 3
Pre-Op Rules: Verify venv (~/.venv) → Run deps check (~/tests/run_tests.sh --deps) → Validate against AGENTS.md → Run test suite → Commit only if passing
Reversibility Policy: >=3 auto-exec; <3 pre-announce (all reversible ops auto, risky ops pre-announced)
Safety Ceiling: Never exceed reversibility floor • All actions git-backed • Pre-op checks mandatory for critical skills
Skill Governance: 12 production skills in ~/.pi/agent/skills/ • All verified by VERIFY_ALL.sh • All tested by unified venv

🤖 From Rules to Practice

Autonomy is not permission—it's a practice. NEXUS inherited governance from AXIOM: written rules (AGENTS.md), safety mechanisms (pre-op checklists, reversibility floors), and enforcement (unified venv, test-first deployment). The system reads these rules on boot and self-checks before every risky action.

📋 The Autonomy Model (AGENTS.md)

AGENTS.md encodes NEXUS's entire autonomy framework:

Three Tiers of Autonomy

Reversibility Floors

Every action is scored on reversibility. If reversibility_floor < 3, the action requires pre-announcement before execution.

Unified Infrastructure

✅ Pre-Operation Checklist (Automated)

Before operating on critical systems (like asciinema-executor or skill deployment), NEXUS runs a strict pre-op sequence:

$ nexus --pre-op-check asciinema-executor
[1/6] Verifying venv exists...
[OK] ~/.venv found • Python 3.14.1
[2/6] Checking dependency cache...
[OK] pytest 9.0.3 • playwright • all deps in ~/.venv/lib
[3/6] Validating autonomy ceiling...
[OK] Operation reversibility: 3 • Allowed at default_level: 3
[4/6] Running VERIFY_ALL.sh for skill...
[OK] asciinema-executor: 7/7 tests passing • Service healthy
[5/6] Starting auto-recovery if needed...
[OK] Service alive • No recovery needed
[6/6] Running core test suite...
[PASS] 156/156 tests passing ✓
[AUTHORIZED] All checks passed • Ready to proceed
$

Only if ALL checks pass does NEXUS proceed. If ANY check fails, the operation is blocked and logged.

🧪 The Unified Testing Framework

Safety emerges from tight feedback loops. The unified venv + test framework creates reproducible validation:

Single Entry Point

Test Matrix (26 Core Tests)

graph TB A["🔧 System Tests
(7)"] --> D["✅ All Pass?"] B["📡 API Tests
(6)"] --> D C["🛠️ Skill Tests
(6)"] --> D D -->|yes| E["🚀 Deploy"] D -->|no| F["❌ Revert"] G["🧠 Memory Tests
(4)"] --> D H["🌐 Endpoint Tests
(3)"] --> D style A fill:#99ff99,stroke:#00ff99,color:#000 style B fill:#99ccff,stroke:#00d9ff,color:#000 style C fill:#ff9999,stroke:#d946a6,color:#000 style G fill:#99ff99,stroke:#00ff99,color:#000 style H fill:#ffff99,stroke:#ffbe0b,color:#000 style D fill:#ffffff,stroke:#00d9ff,color:#000,stroke-width:2px style E fill:#00ff99,stroke:#00ff99,color:#000,stroke-width:2px style F fill:#d946a6,stroke:#d946a6,color:#fff,stroke-width:2px

Test Execution Guarantee

Before any deployment: 156 sub-tests must pass. If ANY fails, the system auto-reverts and escalates. This is not optional—it's embedded in the deployment pipeline.

🛡️ The Philosophy: Safety Over Speed

Autonomy is not recklessness.
NEXUS moves fast, but only on reversible operations.
High-impact or irreversible decisions require human review.

The reversibility floor is the line between agility and safety.

Pre-op checks aren't overhead—they're the price of trust.
When humans can verify every action, agents can move freely within that space.

← GENESIS ⬆ NEXUS HUB DISCOVERY →