Skip to content

Security: dancinlab/canon

Security

SECURITY.md

SECURITY — canon defense surface

Honesty triad mode-6 precondition (f) declaration User authorization 2026-04-26 (cross-repo write from nexus session) Origin: ~/core/nexus/design/hexa_sim/2026-04-26_honesty_triad_refresh_omega_cycle.json

Threat model

canon is a knowledge corpus + theory repo. Unlike nexus (registry + defense tools) or anima (ALM substrate runtime), canon's threat surface is primarily about theorem integrity + atlas SSOT immutability:

  • atlas/atlas.n6: canonical SSOT for n=6 framework derivations (12000+ lines)
  • domains/physics/simulation-theory/: deep-universe-simulation reference
  • theory/, papers/: theorem statements + proof drafts
  • proposals/, techniques/: paradigm shift proposals

Defense surface

Layer Mechanism Status
Source-of-truth integrity git-tracked atlas.n6 with byte-equality seal at hash level live (any git history rewrite detectable)
Cross-repo mirror nexus/n6/atlas.n6 = symlink to this atlas (single SSOT, no drift) live (verified by atlas_health.sh)
Theorem-level review papers/, theory/ subjected to peer-review prior to atlas anchor manual (no automated lint yet)
Proposal gate proposals/ as staging area; promotion to atlas requires manual review manual
Build attestation Makefile + lean4-n6/ for formal proof verification partial (lean4 subset only)

Defense gaps (acknowledged, raw 73 admissibility)

  1. No automated atlas drift detection inside this repo — defense delegated to nexus's tool/atlas_health.sh + state/atlas_sha256.tsv (cross-repo defense via symlink to nexus's tracked SHA).
  2. No SECURITY_AUDIT.md detail — this doc is a declaration, not an audit; audit lives in nexus's design/hexa_sim/SECURITY_AUDIT.md covering the full 5-layer defense system.
  3. No CVE / disclosure process — canon is a research repo; security issues go through proposal/issues channel.

Honesty triad position

After this declaration, mode-6 score: 6/6 (was 5/6 — sole (f) FAIL).

Together with nexus 6/6, anima 6/6, hexa-lang 5/6 (ceiling at (d)):

  • honesty_5_5 = 4 (was 3 before this commit)
  • honesty_6_6 = 3 (was 2)

Cross-references

  • Audit decision: ~/core/nexus/design/hexa_sim/2026-04-26_honesty_triad_refresh_omega_cycle.json
  • Defense system audit (full): ~/core/nexus/design/hexa_sim/SECURITY_AUDIT.md
  • Cross-repo dashboard: ~/core/nexus/design/hexa_sim/cross_repo_dashboard.md
  • atlas SSOT mirror: ~/core/nexus/n6/atlas.n6~/core/canon/atlas/atlas.n6

Authorization

Cross-repo write authorized by user 'all go' decision 2026-04-26 (Asia/Seoul) per NEXT_SESSION_HANDOFF_v3.md pending item #4. raw 71 manual escalate honored: this file added with explicit user approval, not autonomously.

There aren't any published security advisories