* fix(proxy): stream inference responses instead of buffering entire body
The inference.local proxy path called response.bytes().await which
buffered the entire upstream response before sending anything to the
client. For streaming SSE responses this inflated TTFB from sub-second
to the full generation time, causing clients with TTFB timeouts to abort.
Add a streaming proxy variant that returns response headers immediately
and forwards body chunks incrementally using HTTP chunked transfer
encoding. Non-streaming responses and mock routes continue to work
through the existing buffered path.
Closes#260
* docs: update architecture docs and example for inference streaming
* test(example): add direct NVIDIA endpoint tests via L7 TLS intercept
Expand inference example to 4 test cases: inference.local and direct
endpoint, each streaming and non-streaming. The direct path exercises
the L7 REST relay (relay_chunked) to verify it already streams
correctly. NVIDIA_API_KEY is picked up from the sandbox env when
started with --provider nvidia.
* feat(policy): add policy recommendation plumbing — denial aggregation, transport, approval pipeline, and mechanistic recommendations
Implement the infrastructure layer for automated policy recommendations (#204):
- Proto: 9 new RPCs and messages for draft policy lifecycle (submit, get, approve, reject, approve-all, edit, undo, clear, history)
- Persistence: SQLite/Postgres migrations and store methods for draft_policy_chunks and denial_summaries tables
- Server: Full gRPC handler implementations with mechanistic mapper that auto-generates NetworkPolicyRule proposals from denial summaries
- Sandbox: DenialAggregator with MPSC channel, deduplication, periodic flush to gateway via SubmitPolicyAnalysis
- CLI: 'openshell draft' subcommand with get/approve/reject/approve-all/undo/clear/history operations
- TUI: Draft recommendations panel accessible from sandbox policy view
- Docs: Architecture documentation in architecture/policy-advisor.md
* feat(policy): add L7-aware mechanistic mapper and policy advisor CTF example
Add L7 rule generation to mechanistic mapper (build_l7_rules,
generalise_path, looks_like_id) with 3 new unit tests. Add
examples/policy-advisor/ with a 7-gate CTF script, restrictive
sandbox policy, and walkthrough README.
* fix(policy): use sandbox name for denial flush and add TUI draft badges
Fix denial aggregator passing sandbox UUID instead of name to
SubmitPolicyAnalysis, which caused 'sandbox not found' errors on
flush. Add notification badges to the TUI sandbox list and detail
header showing pending draft recommendation counts.
* fix(policy): deduplicate draft chunks and tolerate overlapping OPA rules
Skip draft chunk creation when a pending/approved chunk already covers
the same host:port endpoint, preventing duplicate rules across denial
aggregator flush cycles.
Rewrite three OPA complete rules (network_policy_for_request,
matched_network_policy, matched_endpoint_config) to tolerate multiple
matching policies without triggering a "complete rule conflict" error.
network_policy_for_request becomes a boolean, matched_network_policy
uses a set comprehension with min(), and matched_endpoint_config uses
an array comprehension with index-0 selection.
* feat(tui): interactive draft actions, highlight bar, and detail popup
Rework the draft recommendations panel to match the logs UX:
- Highlight bar (green accent + background) instead of arrow marker
- Viewport-aware j/k scrolling with g/G for top/bottom
- Enter opens a full-screen detail popup showing endpoints, binaries,
rationale, security notes, and action hints
Add approve/reject/approve-all draft actions:
- [a] approve selected chunk, [x] reject, [A] approve all pending
- Actions work from both the list view and the detail popup
- gRPC calls run async; result updates status bar and refreshes data
- Nav bar shows all available keybindings
Fix draft count refresh: sandbox_draft_counts now refreshes on every
tick (not just Dashboard), so the detail header badge updates in
real time.
Improve badge labels: show 'N pending' instead of a bare number in
both the dashboard sandbox list and sandbox detail header.
* refactor(policy): DB-level draft chunk dedup with hit counter and timestamps
Replace the in-memory HashSet dedup in SubmitPolicyAnalysis with a
database-level upsert. New denormalized columns on draft_policy_chunks:
- host, port: extracted from proposed_rule at insert time
- hit_count: incremented on conflict (same sandbox + host + port)
- first_seen_ms, last_seen_ms: track when the endpoint was first and
most recently proposed
A partial unique index (WHERE status IN ('pending','approved')) ensures
only one active chunk per endpoint per sandbox; rejected/superseded
chunks don't block new proposals.
Surface hit_count and first/last_seen in:
- CLI: 'openshell draft get' shows 'Hits: N (first ..., last ...)'
- TUI: detail popup shows hits row; list view shows 'Nx' suffix
* fix(policy): optimistic retry on policy version conflicts + structured logging
merge_chunk_into_policy and remove_chunk_from_policy now retry up to 5
times on UNIQUE constraint violations (version conflicts from concurrent
approvals). Each attempt re-reads the latest policy, re-merges the rule,
and increments the version. This eliminates the race condition where
rapid successive approvals would fail with a DB error.
Add structured tracing to all draft action handlers:
- ApproveDraftChunk: logs rule_name, host, port, hit_count before merge
and version + policy_hash after success
- RejectDraftChunk: logs rule_name, host, port, reason
- ApproveAllDraftChunks: logs pending_count at start, per-chunk merge
progress, and final summary with chunks_approved/skipped
- UndoDraftChunk: logs before/after with rule_name and version
- Retry attempts log as warnings with attempt number and conflicting
version
* wip: forward proxy fix, mapper allowed_ips, TUI polish, CTF rewrite
* fix(tui): use correct --gateway flag for ssh-proxy ProxyCommand
* chore: add Docker cleanup script for stale images, volumes, and build cache
* feat(tui): approve-all confirmation modal and CTF cleanup
Add [A] confirmation popup that snapshots pending chunks, shows a
scrollable list, and approves each chunk individually on confirm.
This prevents approving chunks that arrived after the modal opened.
Remove transient issue #205 reference from CTF victory banner.
* fix(tui): correct import ordering for rustfmt
* wip: stateful toggle model, rename to network rules
Draft chunks now follow a toggle state machine:
pending -> approved | rejected (initial decision)
approved <-> rejected (toggle)
One row per (sandbox_id, host, port) via expanded unique index.
Rejecting an approved rule removes it from the active policy.
Re-approving a rejected rule merges it back.
Rename CLI from 'draft' to 'rule', TUI from 'Draft Recommendations'
to 'Network Rules'. State-aware keybindings: approved shows [x] Revoke,
rejected shows [a] Approve. Fix sandbox detail hiding delete confirmation
behind pending message.
* refactor(policy): move mapper sandbox-side, slim schema, per-binary granularity
Move mechanistic mapper from gateway to sandbox so all analysis runs
sandbox-side (N sandboxes = N independent pipelines). Gateway is now a
thin validate + persist + approval layer.
Architectural changes:
- Move mechanistic_mapper.rs from navigator-server to navigator-sandbox
- Sandbox flush flow: aggregator drains -> mapper runs -> proposals sent
- Gateway SubmitPolicyAnalysis: validate + persist only, no mapper
- Drop denial_summaries table (write-only, zero readers)
- Consolidate migrations 003+004+005 into single 003
Schema slimming:
- Drop 5 unused columns from draft_policy_chunks (stage, denial_refs,
supersedes_chunk_id, analysis_mode, decided_by)
- Add per-binary granularity: binary column, widen unique index to
(sandbox_id, host, port, binary)
- Mapper groups by (host, port, binary), one proposal per triple
- Merge appends binary to existing rule; revoke removes just that binary
CTF & UX:
- 7-gate CTF: add Gate 3 (curl -> ifconfig.me:80) for per-binary demo
- TUI shows binary short name in list, full path in detail popup
- CLI output shows binary field
- Idempotent rule names, hit_count accumulates real denial counts
- Rationale text no longer bakes in stale denial count
Closes#32
Add post-condition checks in drop_privileges() to verify that setgid()
and setuid() actually changed the effective IDs. Also verify that
setuid(0) fails after dropping privileges, confirming root cannot be
re-acquired. This is a defense-in-depth hardening measure per CWE-250
and CERT POS37-C. All added syscalls (geteuid, getegid, setuid) are
async-signal-safe, so they are safe in the pre_exec context.
Co-authored-by: John Myers <johntmyers@users.noreply.github.com>
The sandbox proxy had a control_plane_endpoints allowlist that let
sandboxed processes reach the gateway without OPA evaluation or SSRF
checks. This is a security issue: if the sandbox can talk to the
control plane, it could modify its own policy, exfiltrate data through
provider environment endpoints, or abuse other gateway APIs.
All connections now go through OPA policy evaluation and SSRF
protection with no exceptions. Removes AllowedEndpoint, parse_endpoint_url,
the is_control_plane bypass in handle_tcp_connection, and the engine field
from ConnectDecision.
Closes#42
- Change non-CONNECT proxy response from 405 to 403 to align with
how CONNECT denials are surfaced
- Add structured deny logging for non-CONNECT requests with hostname
extraction from absolute-form URIs
- Revise 7 user-facing error messages across proxy.rs and
dev-sandbox-policy.rego to follow consistent principle: generic
policy-deny messages for non-inference requests, descriptive messages
for recognized inference endpoints
- Update E2E test assertion to match new error message
- Update architecture docs to reflect new behavior
Co-authored-by: John Myers <johntmyers@users.noreply.github.com>
* fix(router): use protocol-aware auth headers for inference proxying
The router always sent `Authorization: Bearer` when proxying to backends,
which breaks Anthropic endpoints that require `x-api-key`. Thread the
source_protocol through to proxy_to_backend and set the appropriate auth
header based on protocol family.
* fix(server): remove noisy inference bundle log
The "serving inference bundle" info log fired every 30s per sandbox on
each route refresh poll. Remove it since there is no equivalent log for
policy serving.
* fix(sandbox): keep inference routing active when cluster bundle is initially empty
In cluster mode, if GetSandboxInferenceBundle returned zero routes at
sandbox startup, build_inference_context() returned None, disabling
inference routing for the sandbox lifetime and preventing
spawn_route_refresh() from starting. Routes created after sandbox
startup were never picked up.
Keep inference interception active with an empty cache in cluster mode
so background refresh can activate routes created later. File-mode
semantics are unchanged.
Closes#78
## What You Can Do
### Live policy updates
Update a running sandbox's network policy without recreating it:
```bash
nav sandbox policy set <sandbox> --policy new-policy.yaml --wait --timeout 60
```
The sandbox hot-reloads the policy within 30s (configurable). On failure, the previous policy stays active (last-known-good).
Idempotent — submitting the same policy twice returns the existing version:
```
✓ Policy version 3 submitted (hash: a1b2c3d4e5f6)
$ nav sandbox policy set test --policy same.yaml
· Policy unchanged (version 3, hash: a1b2c3d4e5f6)
```
### Policy history & inspection
```bash
nav sandbox policy list <sandbox> # version history with status
nav sandbox policy get <sandbox> # current policy metadata
nav sandbox policy get <sandbox> --full # print full policy as YAML
nav sandbox policy get <sandbox> --rev 2 --full # specific revision as YAML
```
### Sandbox logs
Stream logs from both the gateway and sandbox supervisor in one view:
```bash
nav sandbox logs <sandbox> # one-shot, last 2000 lines
nav sandbox logs <sandbox> --tail # live streaming
nav sandbox logs <sandbox> --source sandbox # supervisor logs only
nav sandbox logs <sandbox> --source gateway # gateway logs only
nav sandbox logs <sandbox> --level warn # warnings and errors only
nav sandbox logs <sandbox> --since 5m # last 5 minutes
nav sandbox logs <sandbox> --tail --source sandbox --level info
```
Each log line is tagged with its source and includes structured fields:
```
[1772055394.673] [sandbox] [INFO ] [navigator_sandbox::proxy] CONNECT action=allow dst_host=api.anthropic.com dst_port=443 policy=claude_code
[1772055061.005] [gateway] [INFO ] [navigator_server::grpc] GetSandboxPolicy served from policy history
```
---
## Implementation
### Proto changes
- 7 new RPCs: `UpdateSandboxPolicy`, `GetSandboxPolicyStatus`, `ListSandboxPolicies`, `ReportPolicyStatus`, `GetSandboxLogs`, `PushSandboxLogs` (client-streaming)
- `SandboxLogLine`: added `source` (gateway/sandbox), `fields` (structured key-value map)
- `WatchSandboxRequest`: added `log_since_ms`, `log_sources`, `log_min_level`
- `GetSandboxLogsRequest`: added `sources`, `min_level`
- `PolicyStatus` enum, `SandboxPolicyRevision` message
- `Sandbox.current_policy_version` field
### Server
- **Policy persistence**: New `sandbox_policies` table (SQLite + Postgres) with per-sandbox monotonic versions, status tracking, and policy hash
- **UpdateSandboxPolicy**: Validates static field immutability (filesystem/landlock/process), network mode consistency (Block↔Proxy), deterministic hash comparison for idempotent updates
- **Lazy backfill**: First `GetSandboxPolicy` call creates version 1 from `spec.policy` for existing sandboxes
- **Log broker**: `TracingLogBus::publish_external()` injects sandbox-pushed logs into the same broadcast channel + tail buffer (2000 lines). Server forces `source="sandbox"` and `sandbox_id` on all pushed logs
- **Source/level filtering**: Applied server-side in both `GetSandboxLogs` and `WatchSandbox` streams
- **Version supersession**: When a new version is loaded, all older pending+loaded versions are marked superseded
### Sandbox
- **`OpaEngine::reload_from_proto()`**: Full `from_proto()` pipeline (L7 validation, access preset expansion) with atomic engine swap. On failure, previous engine untouched (LKG)
- **Policy poll loop**: Background task polls every 30s (configurable via `NAVIGATOR_POLICY_POLL_INTERVAL_SECS`), reports status via `ReportPolicyStatus` RPC
- **`LogPushLayer`**: Tracing layer captures events at INFO+ (configurable via `NAVIGATOR_LOG_PUSH_LEVEL`), sends structured fields via `PushSandboxLogs` client-streaming RPC. Background task batches 50 lines / flushes every 500ms. Best-effort (drops on full channel, never blocks)
- **`CachedNavigatorClient`**: Persistent mTLS channel for both policy polling and log push
### Database migration
- `002_create_sandbox_policies.sql` (SQLite + Postgres)
## Tests
- **Unit**: 8 policy persistence tests (put/get/list/status/supersede/isolation)
- **Integration**: 4 test files updated with new RPC stubs
- **E2E**: `test_live_policy_update_and_logs` — full lifecycle: create → set same (unchanged) → push new → wait for load → verify connectivity → push same (unchanged) → verify history → fetch logs
## Documentation
- `architecture/sandbox.md`: Log streaming architecture, LogPushLayer, push task, server broker, source tagging, structured fields, CLI filtering, failure modes
- `architecture/security-policy.md`: Live update semantics, deterministic hashing, CLI filter flags, policy inspection
- `architecture/plans/issue-78-sandbox-log-streaming.md`: Design plan for log streaming
## Security
- Trust boundary documented: shared mTLS cert model (per-sandbox auth tracked in #80)
- Server forces `source="sandbox"` and `sandbox_id` on pushed logs (can't impersonate gateway or other sandboxes)
- Per-batch line cap (100) prevents flooding
> **🏗️ build-from-issue-agent**
Closes#81
## Summary
The proxy was forwarding the client's original `model` field in the request body unchanged to the backend, causing backends to reject requests for models they don't serve. Now `proxy_to_backend()` always sets the `"model"` field in the JSON body to the route's configured model before forwarding. This covers all 4 supported inference protocols (openai_chat_completions, openai_completions, openai_responses, anthropic_messages).
## Changes Made
- `crates/navigator-router/src/backend.rs`: Added JSON body model override — parses the body and always sets `"model"` to `route.model` (replaces if present, inserts if absent). Non-JSON bodies pass through unchanged.
- `crates/navigator-router/tests/backend_integration.rs`: Added 2 integration tests using wiremock's `body_partial_json` matcher
## Deviations from Plan
Minor improvement: the plan specified only replacing `"model"` when already present. Changed to always set it, since all inference protocols require the field and omitting it would cause backend rejection.
## Tests Added
- **Unit:** N/A
- **Integration:** `proxy_overrides_model_in_request_body` and `proxy_inserts_model_when_absent_from_body` in `crates/navigator-router/tests/backend_integration.rs`
- **E2E:** N/A
## Documentation Updated
- `architecture/sandbox.md`: Updated proxy forwarding description to include model ID rewriting
- `architecture/inference-routing.md`: Updated backend.rs description, sequence diagram, and backend proxying steps
## Verification
- [x] All tests passing
- [x] Pre-commit checks passing (clippy, fmt)
- [x] Architecture documentation updated
> **🏗️ build-from-issue-agent**
Closes#79
## Summary
Replaces the gateway-proxied inference model (`ProxyInference` RPC) with sandbox-local execution. The sandbox now resolves routes from either a standalone YAML route file or a cluster bundle fetched via the new `GetSandboxInferenceBundle` RPC, then forwards requests directly to inference backends using `navigator-router`.
### Benefits of removing the gRPC hop
Moving inference execution from the gateway to the sandbox eliminates the gRPC round-trip for every inference request:
- **No payload size limits**: Direct HTTP forwarding replaces protobuf-over-gRPC serialization.
- **Streaming-ready**: Preserves connection semantics for future SSE support (e.g., `stream: true`).
- **Lower latency**: Direct sandbox → backend path instead of sandbox → gateway → backend round-trip.
- **Reduced gateway load**: Gateway only serves the control-plane bundle delivery RPC.
- **Simpler error model**: HTTP status codes flow through directly without gRPC wrapping.
## Changes Made
- **Proto**: Removed `ProxyInference` RPC. Added `GetSandboxInferenceBundle` RPC for route bundle delivery.
- **Gateway**: Replaced `proxy_inference` with `get_sandbox_inference_bundle`. Removed `Router` and `navigator-router` dependency — gateway is now control-plane only for inference.
- **Router**: Switched config file format from TOML to YAML. Added `Debug` impl that redacts API keys.
- **Sandbox**: New `InferenceContext` with local `Router` + route cache. Routes loaded from `--inference-routes` file (standalone) or cluster bundle via gRPC, with 30s background refresh. New `--inference-routes` / `NAVIGATOR_INFERENCE_ROUTES` CLI arg.
- **Dev sandbox**: Added `inference-routes.yaml` at repo root with default NVIDIA NIM route. `mise run sandbox` now mounts it automatically and supports `-e` flag to forward env vars (e.g., `mise run sandbox`).
- **Example**: Added standalone example (`examples/inference/routes.yaml`) and rewrote README to cover both standalone and cluster workflows.
- **E2E tests**: Updated existing test names/docs. Added tests for Anthropic messages protocol and multi-route policy filtering.
## Tests Added
- **Unit (21 new):** `router_error_to_http` (6 variants), `load_from_file` YAML round-trip (5), `resolve_sandbox_inference_bundle` gRPC handler (6), `build_inference_context` route loading (4)
- **E2E (2 new):** Anthropic messages protocol routing, route filtering by `allowed_routes`
- **Total:** 188 unit tests passing across 3 crates
## Documentation Updated
- `architecture/gateway.md`: Removed router, updated Inference Service
- `architecture/sandbox.md`: Updated orchestration flow, InferenceContext, route loading
- `architecture/inference-routing.md`: Complete rewrite for sandbox-local architecture
## Verification
- [x] All unit tests passing (188 across 3 crates)
- [x] Pre-commit Rust checks passing (fmt, clippy, tests)
- [x] Architecture documentation updated
- [x] E2E tests updated with new test cases
- [x] Standalone example and documentation added
- [x] Default inference-routes.yaml with env passthrough for mise run sandbox
## Why
Sandbox proxy policy evaluation was re-hashing unchanged binaries on every CONNECT request. In practice, this made the warm path much slower than necessary.
## What changed
- `BinaryIdentityCache` now stores a file fingerprint with the cached TOFU hash.
- `verify_or_cache()` returns the cached hash immediately when fingerprint is unchanged.
- If fingerprint changes, it re-hashes and compares to the cached hash to preserve tamper detection.
## Scope (fix only)
This MR intentionally contains only the identity-cache fix and docs for that behavior.
It does **not** include benchmark harness or perf-instrumentation setup changes.
## Benchmark results (same environment)
Pre-fix vs post-fix p50:
- `sandbox_connect_local`: `1784.8 ms` -> `1696.4 ms`
- `sandbox_connect_network_via_proxy`: `5692.9 ms` -> `1860.3 ms`
- Derived proxy-path network overhead: `3908.1 ms` -> `163.9 ms` (**95.8% reduction**)
Additional timing signal from perf logs:
- `PERF_SANDBOX_PROXY_CONNECT decision_ms` p50: `3742 ms` -> `1 ms`
## Validation
- `cargo fmt`
- `cargo test -p navigator-sandbox identity::tests`
Closes#67
## Summary
Implements transparent inference interception and routing for sandboxed AI agents. The sandbox proxy intercepts outbound AI SDK calls (OpenAI, Anthropic) and reroutes them through the gateway to policy-controlled backends — enabling organizations to redirect inference traffic to local or self-hosted models without modifying agent code.
**Decision model** — a tri-state OPA evaluation for every CONNECT request:
1. Binary + endpoint explicitly allowed in `network_policies` → **allow** (pass through)
2. Not explicitly allowed + `inference.allowed_routes` configured → **inspect for inference** (TLS intercept, detect API patterns, route through gateway)
3. Otherwise → **deny**
No endpoint declarations or binary lists needed for inference routing. Just configure `inference.allowed_routes`.
## Key Changes
### Sandbox (interception)
- **OPA policy**: New `network_action` Rego rule with three outcomes (`allow`, `inspect_for_inference`, `deny`). New `NetworkAction` enum replaces `PolicyDecision.allowed` bool for the proxy's main decision path.
- **Proxy**: New `InspectForInference` path — TLS-terminates client, parses HTTP, detects inference API patterns (`POST /v1/chat/completions`, `/v1/completions`, `/v1/messages`), strips auth headers, forwards via gRPC.
- **New module**: `l7/inference.rs` — `InferenceApiPattern`, `detect_inference_pattern()`, HTTP request/response parsing.
- **gRPC client**: New `proxy_inference()` for sandbox→gateway forwarding.
- **Sandbox init**: Creates OPA engine when inference is configured, even without `network_policies`.
### Gateway (dispatch)
- **InferenceService**: `ProxyInference` RPC loads sandbox policy, resolves allowed routes, dispatches to router. Full CRUD for inference routes.
- **Proto**: `InferenceRoute`, `InferenceRouteSpec`, `ProxyInferenceRequest/Response`, Inference gRPC service.
### Router (backend proxying)
- **New crate**: `navigator-router` with `Router`, `proxy_with_candidates()`, protocol-based route selection, backend HTTP proxying with auth header rewriting.
- **Mock support** for testing (`mock://` scheme).
### CLI
- `nav inference create/update/delete/list` commands for route management.
### Python SDK
- Updated protobuf bindings. Removed old `inference.py` client (replaced by transparent interception).
### Documentation
- New `architecture/inference-routing.md` — end-to-end system documentation.
- Updated `architecture/sandbox.md` — proxy, OPA, and source index sections.
- Updated `architecture/README.md` — new subsystem overview and diagram.
## Addendum: Chunked Transfer Compatibility
This branch now also fixes intercepted SDK requests that send chunked request bodies:
- `inspect_for_inference` now accepts `Transfer-Encoding: chunked` and decodes chunked request bodies before forwarding to the gateway
- Removed the prior `411 Length Required` response for chunked intercepted requests
- Added request/response header sanitization for framing and hop-by-hop headers (`content-length`, `transfer-encoding`, `connection`, etc.) to keep forwarded requests and returned responses valid
- Added unit tests for chunked parsing and header sanitization
Note: this improves compatibility for streaming-style SDK request patterns; true token-by-token passthrough response streaming is still a separate follow-up.
## Minimal Policy for Inference Routing
```yaml
inference:
allowed_routes:
- local
```
Any outgoing connection from a binary not explicitly allowed in `network_policies` will be intercepted and checked for inference API patterns.
## Test Plan
- [x] `cargo test --workspace` — all tests pass
- [x] `mise run pre-commit` — all checks pass
- [x] E2E: OpenAI chat completions routed through gateway
- [x] E2E: Anthropic messages routed through gateway
- [x] E2E: Python OpenAI SDK from sandbox (`examples/inference/inference.py`)
- [x] E2E test: `e2e/python/test_inference_routing.py`
> **🔧 security-fix-agent**
Closes#62
## Security Fix
### Summary
The CONNECT proxy accepted hostnames from clients and connected to whatever IP they resolved to, with no validation against internal address ranges. While the OPA policy is default-deny, a misconfigured or overly permissive policy could allow SSRF to cloud metadata (169.254.169.254), localhost, or RFC1918 services. This fix adds DNS resolution before connecting and rejects any host that resolves to an internal IP.
### Severity Assessment
- **Impact:** Medium — if exploited, could reach cloud metadata (IAM creds), cluster-internal services, or host-local services
- **Exploitability:** Very low — requires OPA policy misconfiguration or DNS rebinding attack
- **Affected components:** `crates/navigator-sandbox/src/proxy.rs` — `handle_tcp_connection`
### Changes Made
- `crates/navigator-sandbox/src/proxy.rs`: Added `is_internal_ip()` helper that checks IPv4 loopback/private/link-local, IPv6 loopback/link-local, and IPv4-mapped IPv6. Added `resolve_and_reject_internal()` that resolves DNS and rejects internal IPs. Inserted check between OPA allow and `TcpStream::connect`, with control plane endpoints exempt.
- `architecture/security-policy.md`: Added SSRF Protection section with blocked ranges table and flow diagram
- `architecture/sandbox.md`: Updated proxy connection flow diagram and added SSRF protection subsection
- `architecture/README.md`: Added internal IP rejection step to proxy description
### Tests Added
- **Unit:** 17 tests in `proxy::tests` — covers IPv4 loopback/private/link-local, IPv6 loopback/link-local, IPv4-mapped IPv6, public IPs, DNS resolution of localhost/127.0.0.1/169.254.169.254, and DNS failure handling
- **Integration/E2E:** N/A — the proxy runs inside a Linux network namespace; unit tests for IP checking and DNS resolution cover the security boundary
### Documentation Updated
- `architecture/security-policy.md`: New SSRF Protection section with blocked IP ranges and Mermaid flowchart
- `architecture/sandbox.md`: Updated proxy flow diagram and added SSRF protection subsection
- `architecture/README.md`: Added step 4 to proxy description
### Verification
All 85 sandbox tests pass including 17 new proxy SSRF tests. Pre-commit (fmt, clippy, full test suite) passes clean with zero warnings.
Add two sub-agents that have been useful thus far. The documentation agent has memory that is tracked.
Used the docs sub-agent to create detailed docs for all sub-systems + and overview doc.
Closes#26
## Summary
Consolidates sandbox policy data into `dev-sandbox-policy.yaml` and deletes `dev-sandbox-policy-data.rego`. The YAML file becomes the single source of truth for sandbox policy data across both paths:
- **`nav create sandbox`** (CLI reads YAML → proto → gRPC → server stores → sandbox fetches)
- **`mise run sandbox`** (sandbox reads YAML directly via regorus)
## Why YAML
The policy data file is user-facing — it's the interface users interact with to configure sandbox security policy (allowed endpoints, binary identities, filesystem access, etc.).
Keeping this as YAML rather than rego:
- **Decouples data from enforcement engine** — if we change from OPA/rego to a different policy engine, the user's data file stays the same
- **Still works with regorus** — regorus can natively load YAML data via its `yaml` feature, so there's no conversion step in the local dev path
- **Enables future indirection** — we can add a transformation layer between the YAML file and what the policy engine receives, without changing the user-facing format
## Changes
### Policy data (`dev-sandbox-policy.yaml`)
- Flat top-level structure: `filesystem_policy`, `landlock`, `process`, `inference`, `network_policies`
- Renamed `filesystem` → `filesystem_policy` to match rego/proto convention
- Deleted `dev-sandbox-policy-data.rego` (was the rego equivalent of this data)
### Rego rules (`dev-sandbox-policy.rego`)
- All data references updated: `data.sandbox.*` → `data.*` (dropped `sandbox.` prefix)
- `package navigator.sandbox` unchanged (that's the rego rule namespace, not the data path)
### Sandbox entrypoint (`navigator-sandbox`)
- `--rego-policy`/`--rego-data` → `--policy-rules`/`--policy-data`
- `NAVIGATOR_REGO_POLICY`/`NAVIGATOR_REGO_DATA` → `NAVIGATOR_POLICY_RULES`/`NAVIGATOR_POLICY_DATA`
- `from_files()`: loads YAML natively via `regorus::Value::from_yaml_file()`
- `from_strings()` and `reload()`: accept YAML data via `Value::from_yaml_str()`
- `proto_to_opa_data_json()`: emits data at root level (no `"sandbox"` wrapper)
- Enabled regorus `yaml` feature
### CLI (`navigator-cli`)
- `DevSandboxPolicyFile.filesystem` → `filesystem_policy`
### Tests
- All inline test data converted from rego/JSON to YAML
- `test_engine()` uses `from_strings(rules, yaml)` with the dev YAML file
### Documentation
- Updated `architecture/plans/sandbox-policy-refactor/plan.md`, `architecture/plans/opa/plan.md`, `architecture/sandbox.md`, `architecture/containers.md`
Closes#24
## Problem
SSH sessions into the sandbox were **not entering the network namespace** or receiving proxy environment variables. This meant every command run via SSH (the only user-facing path) had unrestricted internet access, completely bypassing OPA network policy enforcement.
The root cause: the SSH server was started **before** the network namespace and proxy were created in `lib.rs`, so it never received the netns fd or proxy URL.
Additionally, **gRPC inference from within the sandbox did not work at all** — even after fixing the netns, multiple issues prevented the Python SDK from reaching the navigator server through the CONNECT proxy.
## Changes
### Sandbox binary
**Core fix — reorder startup + thread netns through SSH:**
- `lib.rs`: Move netns + proxy creation before SSH server start. Compute `ssh_netns_fd` and `ssh_proxy_url`, pass them to `run_ssh_server()`.
- `ssh.rs`: Thread `netns_fd` and `proxy_url` through the full SSH call chain into `spawn_pty_shell()`. Set proxy env vars on the shell command. Call `setns(fd, CLONE_NEWNET)` in `install_pre_exec()`.
**Proxy — control plane allowlist + IPv6 socket lookup:**
- `proxy.rs`: Connections to the navigator endpoint (derived from `NAVIGATOR_ENDPOINT`) are always allowed without OPA evaluation, logged with `engine=control_plane`. This is infrastructure the sandbox needs to function, not a user-configurable policy.
- `procfs.rs`: Extended `parse_proc_net_tcp` to also check `/proc/<pid>/net/tcp6`. gRPC C-core uses `AF_INET6` sockets even for IPv4 connections, so its TCP entries were invisible to the proxy's identity resolver. Also fixed port parsing to use `rsplit_once(':')` for correct IPv6 address handling.
**Proxy env vars — lowercase variants for gRPC C-core:**
- `process.rs` + `ssh.rs`: Added lowercase `http_proxy`, `https_proxy`, `grpc_proxy` alongside uppercase. gRPC C-core (libgrpc) checks lowercase first and was ignoring the uppercase-only vars.
### Server
- `sandbox/mod.rs`: Added `CAP_SYS_PTRACE` to sandbox pod security context. Required for the proxy (root) to read `/proc/<pid>/fd/` of sandbox-user processes for binary identity resolution.
### Python SDK
- `inference.py`: Strip `http://`/`https://` scheme from endpoint before passing to `grpc.insecure_channel()`, which expects `host:port`. Default `endpoint` and `sandbox_id` from `NAVIGATOR_ENDPOINT` / `NAVIGATOR_SANDBOX_ID` env vars so `Inference()` works inside sandboxes with no arguments.
### Build / infra
- `ci.toml`: Added `--cap-add=SYS_PTRACE` to `mise run sandbox` to mirror the k8s pod capabilities.
### Documentation
- `architecture/sandbox.md`: Documented `CAP_SYS_PTRACE` requirement and the full set of proxy env vars (uppercase + lowercase).
## Testing
All 29 sandbox unit tests pass. `mise run pre-commit` passes (fmt, clippy, all workspace tests, python lint).
E2E verified on a live cluster:
| Test | Result |
|------|--------|
| Proxy env vars (6 vars, upper+lowercase) | PASS |
| Blocked endpoints (google, anthropic via curl) | PASS — all denied |
| **gRPC inference from SSH session** (`Inference()` with no args, env var defaults) | **PASS** |
| Proxy log: navigator requests show `engine=control_plane` | PASS |
| Proxy log: blocked requests show `engine=opa` with correct deny reasons | PASS |
Piotr Mlocek