HiveCraft

How practice grows

From early sprout to established plant.

Practice grows in stages. The level-ladder names those stages without arithmetic — one observable scene per stage, so you can read where your own work sits without needing the scoring rules.

How the stages connect

Each stage sits on the one below. A higher stage is not better in the abstract — it is what the practice looks like once a few specific things become reliable. Below: a horizontal walk through the stages, with a short scene at each.

Level timeline

  1. needs care first

    Hindering Practice

    A practice in L0 is doing more harm than good according to its evidence: at least one Class-A (Active-harm) blocker item is failing the evidence test. Signed-score sits at -3 to -1.

    Show details

    Code: L0

    Threshold: Active-harm evidence on Class-A blocker items.

    Group-floor: —

    ⚑ Class-A blockers cannot be waived; remediation is required to leave L0.

  2. Initial / Ad-hoc

    Work happens, but evidence is anecdotal. Hive may be productive day-to-day yet cannot demonstrate repeatability on demand.

    Show details

    Code: L1

    Threshold: Some practice exists, mostly tribal, sporadically documented.

    Group-floor: —

  3. Repeatable

    A baseline of evidence exists. Practices are followable by a second operator without insider context. Class-S (Statutory) blockers must be passing.

    Show details

    Code: L2

    Threshold: Documented practice, executed on ≥2 cycles with retained evidence.

    Group-floor: —

  4. Defined

    A common process is in use across roles and projects. Lessons feed back into the canonical KB. Group B and D group-floors begin to bind.

    Show details

    Code: L3

    Threshold: Standardised across hive; tailoring rules documented; Class-M (Maintenance) blockers satisfied.

    Group-floor: Group D ≥ L3 floor active

  5. Managed

    Telemetry drives decisions. Cohort comparison is meaningful (tier-A or tier-B). Weakest-link rule binds: any dimension below L3 caps the overall claim.

    Show details

    Code: L4

    Threshold: Quantitatively managed; cohort-tier-A or -B comparison; weakest dimension ≥ L3.

    Group-floor: Groups A and C ≥ L4 floor active

  6. Optimising

    Hive operates a closed-loop improvement engine with live-review queries (OD2), an active consent-versioning regime, and audit-invocations on cadence.

    Show details

    Code: L5

    Threshold: Continuous improvement evidence; live-review queue active; full anti-gaming compliance.

    Group-floor: All groups ≥ L4 floor; weakest dimension ≥ L4

  7. horizon

    Self-organising Hive

    L6 is reserved as a horizon target. The HiveCraft canon explicitly disclaims L6 as pre-empirical: it has no calibrated cohort, and any claim must be marked pre-empirical in the Profile JSON header.

    Show details

    Code: L6

    Threshold: Pre-empirical aspirational level — no published L6 evaluations exist.

    Group-floor: All groups ≥ L5; weakest dimension ≥ L5; OSC-style distributed authority pattern.

    ⚑ No practice may publish an L6 rating without pre-empirical disclaimer per §11.

About the highest stage

The highest stage is a stated direction-of-travel rather than a measured certification — there is no calibrated cohort there yet. Any practice claiming it must say so plainly, and any aggregator carrying that claim inherits the same obligation.

Want to see what stage your practice is at?

The self-score will name the stage and the area most worth attention next, without arithmetic.

Score your own practice See hive-types