Handbook

Autonomy levels

Autonomy is not binary. Levels are theoretically unbounded; this ladder gives a meaningful, testable progression for governed agent execution. Each level is defined by its unit of autonomous delivery, what stays fixed…

Introduction

Autonomy is not binary. Levels are theoretically unbounded; this ladder gives a meaningful, testable progression for governed agent execution. Each level is defined by its unit of autonomous delivery, what stays fixed, and where humans gate.

Use the ladder when scoping unattended runs, PDCA campaigns, wizard sessions, or Dark Factory targets—so teams do not claim more autonomy than their gates and resources support.

Implementation readiness (summary)

Operator hub and per-level reference architecture (building blocks, evidence, vision): Forge Platform autonomy levels.

Level Policy PoC demonstrated
L0 Defined N/A (interactive)
L1 Defined Yes
L2 Defined Yes
L3 Defined Yes
L4–L8 Defined (vision gates) No

Full readiness matrix, Wizard ↔ execution mapping, and per-level architecture pages live on the Platform hub — this page remains the canonical policy table.

L0–L8 ladder

Level Autonomous unit What stays fixed Human gate
L0 Assisted Suggestions only Everything Continuous
L1 Function One method/function to a given signature/contract Architecture, API, tests Approve branch/merge
L2 Change-set Multi-function / multi-file defect fix or small change, no rearchitecture Architecture, public contracts Accept acceptance criteria + merge
L3 Use-case slice End-to-end user-visible flow inside one existing app (UI + logic + data + tests) Existing architecture, single platform Intent + acceptance in; review out
L4 Feature/component Capability across modules; may add a component within existing architecture; cross-repo, one platform Platform, major architecture ADR + release gate
L5 Subsystem w/ local arch evolution Introduces patterns/refactors within a platform Platform boundary Architecture decision escalation
L6 Product increment Multi-repo, multi-service on a single cloud/platform Cloud/platform choice Go/no-go
L7 Multi-cloud / multi-platform solution (max − 1) A whole engineered solution spanning clouds/platforms Business framing Strategic checkpoints
L8 (max) Autonomous problem solving Frames a business/humanity problem and composes L7 solutions as puzzle pieces Nothing but the goal Mission definition only

Autonomy ladder L0 to L8 with L0 to L3 emphasized as defined and exercised and L4 to L8 muted as future vision

L0–L3 are defined and exercised today (L1–L3 are demonstrated in the worked examples); L4–L8 remain vision requiring ADRs, go/no-go, and strategic checkpoints. Higher levels add gates and never remove lower-level ones.

How a level is enforced

  1. A run (or campaign item) declares its target level.
  2. The Assay gate for that level must pass before the change is considered done. Core evidence in forge/forge.config.yaml includes tests_pass, acceptance_criteria_met, and risks_reviewed.
  3. Higher levels do not skip lower-level gates; they add gates (for example ADR + release at L4).
  4. At L2, multi-file work should produce proof that two or more distinct files changed when acceptance criteria require a change-set—not a single-file patch dressed as L2.

Resource honesty (local-first)

Fully cloud-free autonomy above L1 is not realistic on a ~4GB local model profile: planning, architecture, and ambiguity exceed small-model capability.

Level band Realistic local-first posture
L0–L1 Achievable with deterministic routing + local worker + verify/repair
L2–L3 Often needs ROI-gated escalation to a larger model or human at pivots
L4+ Requires explicit human gates (ADR, go/no-go, strategic checkpoints) regardless of model

The realistic operating mode is local-first with ROI-gated escalation. Track escalation rate over time; it should fall as capability cards and deterministic scaffolds improve. See Respecting resources.

Wizard alignment (planning intent)

The Blueprints Wizard in Forge Lenses captures planning-time autonomy separately from runtime loops. The AutonomyLevel enum includes l0_analyst, l1_drafter, l2_stage_autopilot, and l3_goal_autopilot, with MutationPolicy describing how far downstream automation may edit artifacts.

Wizard policies inform prompts and downstream automation; they do not silently apply upstream edits. Persisted session policies should match the ladder level you intend for execution. Full Wizard ↔ execution mapping: Platform autonomy levels hub.

Forge Dark Factory (PoC reference implementation)

Forge Dark Factory is the current PoC reference implementation for bounded autonomous coding—not a production Platform submodule.

Forge Dark Factory bounded execution loop: classify, route, context, plan, draft, apply, verify, proof, trace, escalate

The governed loop for L1–L3. Verify failures trigger bounded repair; ambiguity or budget exhaustion escalates to a human, who still approves the branch or merge.

Aspect PoC scope today
Target autonomy L1–L3 demonstrated (function, change-set, use-case slice); see worked examples
Loop Classify → route → context → plan → draft → apply → verify → repair → proof → dual-wiki trace → escalate
Dependencies forge-lcdl (patch units, verify, repair, proof); forge-workcells (optional local worker)
Trace Machine record (M) + generated human narrative (H) with freeze gate
Routing Deterministic Cynefin × t-shirt × value; decompose before cloud/human escalate

Do not treat Dark Factory as compliance-ready or as permission for unsupervised push/deploy. It demonstrates how the ladder and respecting resources rules compose in code.

What we do not claim

  • No unsupervised push/deploy — Git workflow and release decisions remain human-gated unless your org explicitly automates them with separate policy.
  • No compliance-ready autonomy — The ladder is engineering governance, not a certification.
  • No “fully autonomous” delivery — Even L8 assumes mission definition by humans; intermediate levels add explicit gates.
  • Escalation is expected — Especially for architecture, security, and ambiguous work; a low escalation rate is a goal, not a guarantee on day one.