Home

Privacy & Sandboxing — Security Architecture Specification

Formal description of Hydite IDE client-side privacy tiers and local sandboxing


Scope: technical alignment document for security audits, compliance assessment, and enterprise procurement review. Not a substitute for the customer's internal risk assessment.

Conformance language: MUST / MUST NOT / SHOULD / MAY follow RFC 2119.

1. Scope and terminology#

1.1 In scope#

  • Data residency (source code, project semantic index, prompt context, model responses, telemetry).
  • Local process sandboxing (filesystem, commands, network, third-party plugin / MCP / Skill invocation).
  • Availability and switching constraints of the three privacy tiers.
  • Transport and trust relationship between the IDE client and the Hydite Vtslx AO gateway.

Out of scope: server-side governance of the gateway (see Vtslx AO docs); internal controls of customer-deployed local inference services — only the boundary with them is specified.

1.2 Definitions#

TermDefinition
Code AssetSource code, configuration, documentation and derivatives tracked by Git in the workspace.
Project IndexThe code-semantic, symbol, call-graph and related-resource index built locally by Context Hub.
Prompt ContextFull text and structured fields assembled and submitted to the inference backend on a single LLM invocation.
On-device ModelA model whose inference runs entirely on the client host or a private service in its local network — no public-internet requests.
Reachable Cloud ModelAny model reached via the public internet, including SaaS APIs and third-party proxies.
Agent Tool InvocationAny filesystem, command, network, MCP or Skill action performed by an Agent.
Customer-Controlled DomainExecution environments under the customer's administrative control: client host, customer LAN, customer-hosted Vtslx AO gateway.

1.3 Control identifier convention#

CTL-<DOMAIN>-<NN> where <DOMAIN>CFG.

2. Threat model#

Simplified STRIDE, restricted to client-side privacy and sandboxing.

ThreatAssetDesign goal
Information disclosureCode, index, prompt contextMUST NOT leave the customer-controlled domain without explicit authorisation.
Elevation of privilegeClient FS, network, peripheralsAgent tool invocations MUST pass through the sandbox; out-of-scope actions denied and auditable.
TamperingProject files, index, configThird-party Skill / MCP MUST NOT write outside the authorised scope.
Replay & MITMTransport with the gatewayAll cross-network transport MUST use authenticated, encrypted channels.
RepudiationAgent action traceabilitySignificant actions SHOULD leave a locally-readable audit record.
Denial of serviceClient resourcesBounded envelopes; no SLA under extreme attack.

3. Trust boundaries and data classification#

3.1 Boundaries (CTL-BND-01..04)#

1
T0: IDE main process
2
└─ T1: Agent tool sandbox (OS primitives)
3
└─ T2: Command / network sub-boundary
4
└─ T3: Third-party Skill / MCP
5
│ Only when mode allows + user authorises
6
7
Vtslx AO gateway (Shared / Dedicated / Self-Hosted)
IDBoundaryEnforcement
CTL-BND-01T0 IDE processHost OS process isolation
CTL-BND-02T1 Agent sandboxmacOS Seatbelt / Linux namespaces
CTL-BND-03T2 Command/networkFine-grained policy within T1
CTL-BND-04T3 Skill/MCPCapability delegation within T1

3.2 Data assets (CTL-DAT-01..06)#

IDAssetDefault residency
CTL-DAT-01Source / Git-tracked filesCustomer-controlled
CTL-DAT-02Project semantic indexCustomer-controlled (local)
CTL-DAT-03Agent worktree / temp branchCustomer-controlled (local)
CTL-DAT-04Prompt contextDetermined by active privacy mode
CTL-DAT-05Model responsesCustomer-controlled
CTL-DAT-06Client telemetryDetermined by user authorisation

4. Formal privacy-mode matrix#

4.1 Tier definitions (CTL-PRV-01..03)#

IDTierFormal definition
CTL-PRV-01High PrivacyFor all of CTL-DAT-01..05, the asset MUST NOT leave the customer-controlled domain; inference MUST target on-device models only.
CTL-PRV-02StandardDefault tier; minimised prompt context (CTL-DAT-04) MAY be sent to user-authorised reachable cloud models; remaining assets MUST stay in the customer-controlled domain.
CTL-PRV-03Self-Hosted EnterpriseInference MUST traverse a customer-hosted Vtslx AO gateway; no asset MUST transit Hydite-controlled network nodes.

4.2 Behaviour matrix#

Asset ↓ / Tier →High PrivacyStandardSelf-Hosted
CTL-DAT-01 Source codeStays on hostStays on hostStays in customer VPC
CTL-DAT-02 Index (build)LocalLocalLocal
CTL-DAT-02 Index (storage)LocalLocalLocal
CTL-DAT-04 Prompt contextStays on hostMinimum-relevant slices to authorised cloudCustomer gateway only
CTL-DAT-05 Model responseFrom on-device modelFrom authorised cloud modelFrom customer-gateway backend
CTL-DAT-06 TelemetryMUST be offMAY be on (user-controlled)MUST route through gateway or be off

4.3 Mode availability constraints (CTL-PRV-04)#

  • The High Privacy toggle MUST be visible only in Agent mode. In Edit mode — whose scope is bounded to the current selection and which does not autonomously call the cloud — no separate toggle is exposed.
  • When High Privacy is active, the input-box pill MUST show a lock indicator; the status bar SHOULD continuously surface the active privacy level.
  • Any attempt to invoke a reachable cloud model while High Privacy is active MUST be denied. If the user has explicitly allow-listed certain cloud models, invocation MUST require an additional confirmation.

5. Sandbox control family#

Hydite Agent runs inside the T1 boundary. This section defines its control family.

5.1 Filesystem (CTL-SBX-01..04)#

IDControlDefault policy
CTL-SBX-01Workspace read scopeProject root and descendants only
CTL-SBX-02Workspace write scopeProject root only; out-of-project writes MUST be explicitly prompted
CTL-SBX-03OS-sensitive paths~/.ssh, ~/.aws, keychains, etc. MUST require explicit allow-listing
CTL-SBX-04Project file denylistPaths matching .hydite/agent-deny.glob MUST be denied for read/write

5.2 Command execution (CTL-SBX-05..07)#

IDControlDefault policy
CTL-SBX-05Working directoryLocked to project root by default
CTL-SBX-06Network from commandsMAY be globally disabled; MUST be disabled by default under High Privacy
CTL-SBX-07Resource limitsSubprocess wall-clock, CPU, memory SHOULD be subject to configurable caps

5.3 Network (CTL-NET-01..04)#

IDControlDefault policy
CTL-NET-01Default egressHigh Privacy: deny; other tiers: only authorised destinations
CTL-NET-02Inference egressOnly the destinations agreed for the active tier (on-device / authorised cloud / self-hosted gateway)
CTL-NET-03Client update channelSHOULD be disable-able or proxy-able by enterprise admins
CTL-NET-04Third-party MCP egressSame envelope as CTL-NET-01; MCP MUST NOT bypass the sandbox

5.4 Third-party code: Skill / MCP (CTL-SBX-08..10)#

IDControlDefault policy
CTL-SBX-08Skill FS rightsBounded by the parent Agent's T1 limits; cannot escape
CTL-SBX-09MCP resource allowlistMUST be explicitly enabled per project by the user
CTL-SBX-10Skill / MCP egressBounded by CTL-NET-04

5.5 Worktree physical isolation (CTL-SBX-11)#

When Agent Hub runs multiple agents in parallel, each agent MUST execute in its own Git worktree (or temporary branch); writes MUST NOT directly land on the main branch and SHOULD pass a unified review before merging. Consequences:

  • A single agent's failure or out-of-scope attempt MUST be confined to its worktree.
  • Any agent's writes MAY be rolled back by deleting its worktree in one click.

6. High Privacy mode — formal specification#

6.1 Activation conditions (CTL-PRV-05)#

  • The current session MUST be in Agent mode.
  • The user activates the toggle on the input-box bottom-left pill, and MUST clear an explicit confirmation panel.
  • The confirmation panel MUST itemise: source-code egress, index egress, command network, Skill / MCP network, telemetry.

6.2 Egress contract (CTL-PRV-06)#

Once High Privacy is active:

  1. Source code MUST NOT leave the host. Any attempt to send file content, paths or metadata to a non-on-device model MUST be denied.
  2. The project semantic index MUST be built and queried by on-device models only. Index data MUST NOT appear in any outbound request body.
  3. Prompt context MUST be sent only to on-device models.
  4. Telemetry MUST be off.
  5. Network access from commands MUST be denied by default. The user MAY allow specific intranet destinations one rule at a time, but public-internet destinations MUST NOT be allowed.

6.3 Inference path restriction (CTL-PRV-07)#

Under High Privacy, inference is restricted to on-device models — including but not limited to:

  • Local inference frameworks running directly on the client host.
  • Private inference services reachable on the client's local network.

Any inference path pointing to the public internet MUST be disabled. If the user has registered specific cloud models in a permitted-list, invocation MUST trigger an extra confirmation before proceeding.

6.4 UI invariants (CTL-PRV-08)#

  • The input-box pill MUST show a lock indicator, visible only while High Privacy is active.
  • The status bar SHOULD display "Local-only" or a semantically equivalent badge.
  • Exiting High Privacy MUST trigger an explicit prompt that explains the egress policy that will resume.

7. Standard mode — formal specification#

7.1 Data minimisation (CTL-PRV-09)#

  • Prompt context MUST pass through Context Hub's local retrieval phase and be reduced to the slices most relevant to the current request.
  • Whole-directory scans or full-index snapshots MUST NOT be placed directly into prompt context.

7.2 Transport protection (CTL-PRV-10)#

  • Cross-network transport MUST use authenticated, encrypted channels.
  • The client MUST verify peer identity. When configured to use a customer-supplied gateway, the client MUST verify the customer-provided certificate or SPKI anchor.

7.3 User-controllable items (CTL-PRV-11)#

The following items MUST be disable-able under Standard mode:

  • Telemetry reporting.
  • Client auto-update channel.
  • Third-party Skill / MCP egress.
  • Local response caching.

8. Self-Hosted Enterprise mode — formal specification#

8.1 Deployment assumption (CTL-PRV-12)#

  • The customer deploys a Hydite Vtslx AO gateway in their own VPC / intranet.
  • The client MUST be configured to a customer-supplied gateway address; Hydite-controlled network nodes MUST NOT appear in the call chain.

8.2 BYOK and identity (CTL-PRV-13)#

  • Model API keys and inference credentials MUST be held by the customer (in the gateway or a customer KMS).
  • The client MUST support enterprise SSO and SHOULD reuse its session identity for gateway authentication.

8.3 Audit cooperation (CTL-PRV-14)#

  • All cross-network requests MUST traverse the customer gateway, which SHOULD provide complete access logs.
  • The client SHOULD support exporting a local activity summary that can be reconciled with the gateway logs for unified audit.

9. Telemetry and minimisation#

9.1 Telemetry classification (CTL-AUD-01)#

CategoryExamplesHigh PrivacyStandardSelf-Hosted
Crash reportsStack, version, platformOffUser can disableThrough gateway or off
Performance countersStartup time, index time (no paths)OffUser can disableThrough gateway or off
Usage countersMode-switch count, agent-launch countOffUser can disableThrough gateway or off
Prompt / responseModel input and outputMUST be offMUST NOT default to onMUST go through gateway or be off

9.2 Defaults (CTL-AUD-02)#

  • Any telemetry containing code snippets, file paths, prompts, or model responses MUST NOT be enabled by default.
  • Default-enabled telemetry items MUST NOT contain fields from which code assets could be reconstructed.

10. Auditability and self-verification#

10.1 User-level self-verification (CTL-AUD-03)#

Verification targetActionExpected result
Index locationCheck .hydite/index/ in the project rootIndex files exist locally with no remote-sync evidence
Network egressRun a host packet sniffer (e.g. lsof -i, Little Snitch, tcpdump)No public-internet inference traffic during High Privacy
Command networkHave the agent run curl https://example.comDenied by sandbox under High Privacy
File over-readAdd *.pem to .hydite/agent-deny.glob, then have the agent read itRead denied and recorded
Mode UISwitch to High PrivacyLock indicator visible on the input-box pill

10.2 Enterprise-level self-verification (CTL-AUD-04)#

  • The customer's network team MAY configure egress allow-listing for client clusters that only permits the self-hosted Vtslx AO gateway, and SHOULD observe no traffic to unexpected destinations.
  • The customer gateway logs SHOULD reconcile, by request count, with locally exported activity summaries.

11. Configuration surface reference#

11.1 Project files (CTL-CFG-01)#

PathPurpose
.hydite/index.ignoreIndex denylist (.gitignore syntax)
.hydite/index.includeForce-included paths for the index
.hydite/agent-deny.globDenylist for Agent file read/write
.hydite/index/Local index storage (recommended in .gitignore)

11.2 Command palette (CTL-CFG-02)#

CommandPurpose
Hydite: Toggle High Privacy ModeToggle High Privacy under Agent mode
Hydite: Rebuild Project IndexForce-rebuild the local index
Hydite: Open Privacy Scope PanelOpen the High Privacy scope confirmation panel
Hydite: Export Local Activity SummaryExport a local activity summary for audit

11.3 Settings (CTL-CFG-03, excerpt)#

KeyMeaning
hydite.privacy.defaultModeDefault privacy mode at IDE start
hydite.privacy.allowedCloudModelsCloud models permitted under High Privacy
hydite.telemetry.enabledWhether basic telemetry is enabled
hydite.network.allowedHostsEgress allowlist for command execution

12. Conformance levels#

For internal customer assessment, two conformance levels are defined:

  • Baseline conformance: client default-enabled controls satisfy §5 and §9.
  • Enterprise conformance: in addition to baseline, the deployment runs in Self-Hosted Enterprise mode and passes §8 and §10.2 self-verification.

This specification makes no claim of certification under any specific framework (e.g. SOC 2, ISO/IEC 27001, FedRAMP). When deploying in regulated industries, customers SHOULD treat this specification as one input to their compliance self-assessment alongside industry-specific requirements.

13. Revision history#

VersionScope
1.0Initial publication: terminology, trust boundaries, three-tier formal matrix, sandbox control family, verifiable items

14. References#