Instructions to use SZLHOLDINGS/szl-ouroboros with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Kernels
How to use SZLHOLDINGS/szl-ouroboros with Kernels:
# !pip install kernels from kernels import get_kernel kernel = get_kernel("SZLHOLDINGS/szl-ouroboros") - Notebooks
- Google Colab
- Kaggle
szl-ouroboros
The a11oy bounded agent-loop trace and its loop-tax decomposition — as a get_kernel-discoverable, offline, stdlib-only governance kernel. From SZL Holdings.
Honest scope box — read this first
This is NOT a trained model. There are NO weights (
.safetensors/.bin/.pt/.gguf). It is a pure-Python, stdlib-only governance kernel: an offline replay of the live Alloy surface — it reconstructs the a11oy agent-loop trace and its loop-tax decomposition from a run's provider-attempt windows. No network, no torch, no tensors — nothing beyond sums, max and subtraction.get_kernel-discoverable purely so the SZL family loads the same way. Λ is not touched here and stays Conjecture 1 (open).
The whole point: honest MEASURED vs DERIVED labels
Loop timing is easy to overclaim. This kernel labels every field, carrying the backbone's
LOOP_TIMING_BASIS string verbatim:
| Field | Label | Meaning |
|---|---|---|
modelMs |
MEASURED | Σ of every provider-attempt wall window (success AND failed hops occupied the loop) |
peakAttemptMs |
MEASURED | the single slowest attempt window (peak-vs-sum) |
overheadMs |
DERIVED | max(0, wall − modelMs) — Alloy's own orchestration, NOT pure CPU time |
serializationTaxMs |
DERIVED | max(0, modelMs − peak) — a COUNTERFACTUAL, never a realized saving (the loop is strictly sequential) |
deadHopMs |
DERIVED | Σ failed windows before the served hop — upper bound only, no prefetch |
maxBudget |
DECLARED | the per-run attempt ceiling |
exit |
REPORTED | converged / budgetExhausted / aborted / error |
receiptsInEqOut |
DOCTRINE | one receipt trail in, one out — a doctrine invariant, NOT a proof |
When wall_ms=None, overheadMs is honestly None / UNAVAILABLE — never fabricated; the
peak/serialization/deadHop split still holds because it needs only the attempt windows. A wall smaller
than modelMs surfaces wallLessThanModel (a measurement inconsistency made visible, never hidden).
Quickstart (fully offline)
from kernels import get_kernel
ou = get_kernel("SZLHOLDINGS/szl-ouroboros", revision="main", trust_remote_code=True)
attempts = [
{"provider": "sovereign", "model": "own-metal", "ok": False, "latency_ms": 220, "node": "tower"},
{"provider": "sovereign", "model": "own-metal", "ok": True, "latency_ms": 900, "node": "laptop"},
]
trace = ou.build_loop_trace(attempts, wall_ms=1300, exit="converged", max_budget=4)
print(trace["modelMs"], trace["overheadMs"], trace["deadHopMs"]) # 1120 180 220
print(ou.explain(trace["labels"])) # per-field MEASURED/DERIVED provenance
print(ou.selfcheck()) # falsifiable arithmetic self-check
LOOP_DOCTRINE = "bounded, terminating, receipt-closed" — the Ouroboros closes on its own tail.
withinBudget (steps ≤ maxBudget) surfaces the bounded-loop guarantee; a violation is reported,
never clamped. Demo runs make no model call → modelMs / peak / deadHop are honestly 0.
Tests — MEASURED, local, stdlib-only
PYTHONPATH=build/torch-universal python3 -m pytest tests/ -q
13 passed
13/13 passed on a local python3 run (MEASURED, stdlib-only). The arithmetic is falsifiable — a
wrong split flips the asserts; an unmeasured wall yields overheadMs=None (UNAVAILABLE, not fabricated);
a wall < modelMs is flagged; a missing latency raises (never coerced); a budget violation is surfaced.
Provenance
Mirrors the live a11oy backbone loop-tax accounting (artifacts/api-server/src/lib/backbone.ts —
buildLoopTrace, sumAttemptMs, peakAttemptMs, deadHopMs, LOOP_DOCTRINE, LOOP_TIMING_BASIS).
Backed by the Lean 4 formalization szl-holdings/lutar-lean,
DOI 10.5281/zenodo.20434308. Λ uniqueness = Conjecture 1 (open).
Compatibility
Python 3.9+, standard library only — no torch, no third-party deps.
License
Apache-2.0. Copyright 2026 SZL Holdings.
The SZL honest-model trio + the kernel suite
Part of the SZL Holdings honest-model trio — stdlib-only offline replays of the live Alloy surface:
| Model | Lane |
|---|---|
szl-invariants |
8 falsifiable receipt/ledger invariants, offline |
szl-ouroboros (this repo) |
bounded-loop trace + MEASURED/DERIVED loop-tax accounting |
szl-formulas |
the 21 canonical formulas with per-formula PROOF-STATUS |
Anchored by the governed-kernel suite szl-kernels
(hub) and its members (szl-lambda-gate,
szl-governed-norm,
governed-inference-meter).
Live substrate: a-11-oy.com · a11oy Space.
SZL Holdings · Ouroboros = bounded, terminating, receipt-closed · modelMs/peak MEASURED, the rest DERIVED · serializationTax is a counterfactual, never a saving · Λ untouched = Conjecture 1 · a-11-oy.com · github.com/szl-holdings · huggingface.co/SZLHOLDINGS
Signed-off-by: Stephen Lutar stephenlutar2@gmail.com
- Downloads last month
- -