Safety Model

Important

Diagnostics may explain and propose. Mutation must be explicit, structured, validated, and reject uncertainty.

Rendering never modifies source files. Automatic remediation is limited to structured edits that:

1. are marked MachineApplicable;
2. still match the expected source contents;
3. use valid UTF-8 boundaries;
4. don't overlap;
5. satisfy declared preconditions.

🧱 Safeguards

• stale-edit rejection and guarded insertion requirements
• transaction-wide validation, recovery state before writes, rollback on observed write failures
• optional post-fix verification, with rollback when it fails
• trusted-root requirements for hydrated compiler edits
• fail-closed documentation version resolution (ambiguous package versions don't guess)
• revision-aware rendering that fails closed when source identity no longer matches
• suggested shell commands are never executed automatically

🔐 Privacy

• Diagnostic history is privacy-light: no messages, source text, arbitrary attributes or remediation payloads.
• Capsules and exports are privacy-aware; bridge attachments are private unless explicitly opted in.
• Forensic output separates evidence from causation; causal claims stay not established unless explicitly recorded.

🔒 Reporting a vulnerability

Email security@cybercoretech.net and please don't open a public issue.

← 🧾 Schemas & Specs  ·  🎨 Themes & Documentation →