Positioning
Why Publicus is different
AI makes code generation cheap. Security evidence, provenance, policy mapping, and review accountability are still expensive. Publicus exists to compress that verification work into an open, inspectable workflow.
Not a model, not a closed scanner
Frontier cyber models create capability.They can reason over complex cyber tasks, but access, policy, and safety controls can be selective. Publicus stays useful because it is model-agnostic evidence infrastructure.
Package scanners find issues.Publicus adds signed registry decisions, source freshness, review proposals, Canada-context mapping, and exportable release evidence.
SBOM tools inventory software.Publicus turns package inventory into fix-first release review: what blocks, what is unknown, what changed, and what source supports the claim.
Closed compliance products hide method.Publicus publishes the evidence contract, routes, source families, freshness lanes, and governance boundaries so claims can be challenged.
The durable value
When advanced cyber models become broadly available, more people will be able to generate and inspect code quickly. That raises the value of a public trust substrate, because the bottleneck moves from generation to verification.
Upload firstUse package files and SBOMs from AI-built projects instead of asking a broad question first.
Evidence before synthesisRegistry truth, vulnerability evidence, source freshness, and package health stay visible before any recommendation layer.
Canada-ready, not Canada-onlyThe first deep context is Canada: CPCSC, CCCS AI/ML supply-chain guidance, TBS responsible AI, and standards crosswalks. The method works anywhere trust matters.
Open contribution, stewarded publicationAnyone can challenge evidence. Load-bearing registry claims still require trusted review before publication.
Primary references
OpenAI Trusted Access for Cyber DefenseUseful context for why frontier cyber capability is valuable but access-gated.
Anthropic Claude Mythos PreviewUseful context for model capability and controlled release.
Meta Llama safety toolingUseful context for open ecosystem safety tools such as CyberSecEval and Code Shield.
CCCS AI/ML supply-chain guidancePrimary Canada-context guidance for AI supply-chain risk.
CPCSC Level 1Primary Canada-context supplier-assurance reference.