tibet-timevector
Causal TimeVector — Mattern/Fidge vector clocks for TIBET/JIS lanes.
TIBET's Lamport generations give a monotone successor order — enough for causal succession, but they cannot tell, across multiple independent JIS lanes, whether A happened before B, B before A, or A and B were genuinely concurrent (under partition / offline work). For that you need a vector clock (Mattern/Fidge 1988). That's this package.
The vector is causal evidence. Wall-clock time is supporting evidence only. JIS/TIBET signatures remain the authority.
It is a separate layer from tibet-y2k38 (which
adapts broken wall clocks). This is the JIS-causality-flex layer.
Two things it gives you
1. Honest causality across lanes
from tibet_timevector import CausalTimeVector
codex = CausalTimeVector.genesis("codex.aint", "codex_0").tick("codex_1")
claude = CausalTimeVector.genesis("claude.aint", "claude_0").tick("claude_1")
codex.relation_to(claude) # "concurrent" — neither saw the other
claude2 = claude.observe(codex, "claude_saw_codex")
codex.relation_to(claude2) # "before" — claude2 now knows codex_1
Relation rule (Mattern/Fidge): A < B iff every component A[i] <= B[i] and at least one
<; symmetric for >; otherwise concurrent. This makes causal drift visible instead
of pretending timestamps are enough.
2. CausalGuard — refuse causally-spontaneous actions
The sharper use, and the reason it's a security primitive. A sleeper agent (e.g. Red
Specter's DETONATE) breaks cause-and-effect: on its trigger it initiates a destructive
action out of nothing — no cryptographically-proven user-intent, no logical predecessor
in the chain. CausalGuard rejects it not because it looks malicious, but because it is
causally impossible: it has no valid ancestor tracing back to a consented intent.
from tibet_timevector import CausalTimeVector, CausalGuard
guard = CausalGuard()
root = CausalTimeVector.genesis("agent.aint", "grant_root")
guard.register(root, consented_root=True) # e.g. produced by a ParentAttest grant
guard.register(root.tick("legit_action")) # descends from the grant → allowed
# A DETONATE action that arises from nowhere:
sleeper = CausalTimeVector(lane_id="agent.aint", event_id="detonate",
generation=99, vector={"agent.aint": 99},
parent_event_id="evt_that_never_existed")
guard.verify(sleeper).allowed # False — "orphan: parent not in the known chain"
You don't need to recognise the attack. You need to recognise the absence of a valid predecessor. That's a fundamentally stronger model than signature/policy matching.
Rejection classes (all = "causally impossible", not "naughty"):
- spontaneous — no parent link and not a consented root
- orphan — parent_event_id references an event never seen (forged / out-of-nothing)
- broken causality — candidate vector does not dominate its claimed parent
- unrooted — chain does not lead back to any consented root
CLI
pip install tibet-timevector
tibet-timevector demo # two-lane concurrency → observe()
tibet-timevector guard-demo # spontaneous-intent rejection (DETONATE)
tibet-timevector check a.json b.json # relation between two vectors
Where it fits
- Mahipal's drift-test: the vector clock IS the answer to "causal timeline + drift" — it distinguishes concurrency from causality under partition, which Lamport can't.
- W3C / VC over TAT: tells which causal world a Verifiable Credential knew at the moment of transfer/presentation/re-attestation — distinct from a linear verifiable history.
- T-1 genesis + sleeper defense (Principe 3): CausalGuard is the runtime gate that makes a DETONATE-style spontaneous action causally impossible. See the SPECTER SLEEPER defensive mapping for the full vector-by-vector analysis.
Dependencies (bootstrap-or-die)
tibet-core >= 0.5.0b2— provenance chainjis-core >= 0.4.0b1— identity / lane authority
Credits
Reference implementation: Codex (tibet-causal-timevector sketch, 2026-05-31). Productionisation + CausalGuard (spontaneous-intent rejection): Root AI, 2026-06-01. Concept + layer-split (vs tibet-y2k38): Jasper van de Meent.
Part of HumoticaOS. One love, one fAmIly. 💙
Metadata
Release files for tibet-timevector 0.1.0
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | Provenance | Actions |
|---|---|---|---|---|
| tibet_timevector-0.1.0.tar.gz | 9.7 kB | Without provenance | Details for tibet_timevector-0.1.0.tar.gz |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Actions Reset |
|---|---|---|---|---|
| tibet_timevector-0.1.0-py3-none-any.whl | Python 3 | none | any | Details for tibet_timevector-0.1.0-py3-none-any.whl |
Total release size: 19.2 kB
Release files / tibet_timevector-0.1.0.tar.gz
| Download URL | tibet_timevector-0.1.0.tar.gz |
|---|---|
| Size | 9.7 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
5131d14711bb8b02c260e37c0c2d378ea76f33088a3ed263ad86ac61b665334b
|
|
BLAKE2b-256 checksum How to use checksums |
482e57ecea0eb26b0e324d08369e4ba35f2f667f00c48a6c4513f4740a5bb8b7
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/6.2.0 CPython/3.13.5
|
Release files / tibet_timevector-0.1.0-py3-none-any.whl
| Download URL | tibet_timevector-0.1.0-py3-none-any.whl |
|---|---|
| Size | 9.5 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
b656c688f2c7947cfcbe4b5154dad43ecbc507d86a3b79a78a26dd1c019b15a7
|
|
BLAKE2b-256 checksum How to use checksums |
a48e73490360e901e9347d034a00e361b3781921102b9863ec759c45d26ca395
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/6.2.0 CPython/3.13.5
|
Security for tibet-timevector 0.1.0
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
No provenance found
None of the files in this release were uploaded with provenance metadata.
A note on package security
PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.
To learn more about protecting yourself from malicious actors, visit our package security documentation.