Plan safe host-mounted workspaces (4.2.2)

[lodyai]

Summary

This draft PR carries the ExecPlan for roadmap step 4.2.2, "Implement safe
host-mounted workspaces", at docs/execplans/4-2-2-safe-host-mounted-workspaces.md.
The plan materialises host-mounted workspaces with an explicit deny-by-default
allowlist, symlink-resolving canonicalisation, symlink-escape rejection, and a
rootless write-permission probe, then attaches a hardened typed bind mount.

• Hexagonal split: pure domain + planner in src/api/host_mount.rs, a
  single-pass HostPathProbe driven port, and a default adapter in
  src/engine/connection/host_mount/. The engine adapter populates
  HostConfig.mounts with a typed bollard::models::Mount (RPRIVATE
  propagation, non_recursive, explicit read_only) only when
  workspace.source = "host_mount".
• New [mounts] configuration section and two PODBOT_MOUNTS_* environment
  variables; additive FilesystemError variants for allowlist and
  write-probe failures; rstest / rstest-bdd / proptest /
  testcontainers coverage.

The plan is in DRAFT and must be approved before implementation begins.

Revision (2026-06-18)

This revision supersedes the original 2026-05-29 draft on this branch after a
fresh round of external research and a community-of-experts design review. Key
upgrades: the writability probe now runs through a single held cap-std
capability (closing the in-process resolve-then-probe TOCTOU window); a
CanonicalHostPath newtype makes non-canonical containment checks
unrepresentable; the residual cross-process mount-time race (runc
CVE-2025-31133 class) and the operator's trusted-parent-chain precondition are
stated precisely and embedded as a verbatim threat-model boundary; the
negative-coverage matrix gains the sibling-prefix trap, exact-root match,
symlinked-root, and no-litter cases; googletest, pretty_assertions, an
insta denial-message snapshot, and a trybuild signature-stability fixture
are adopted; and the choice of proptest over kani/verus is justified.
See the plan's Supersession note and Decision log for the full rationale.

Test plan

• make check-fmt
• make lint
• make test (rstest + rstest-bdd + proptest)
• testcontainers-backed end-to-end scenario for the host-mount bind
• make markdownlint, make nixie

References

• Roadmap entry: docs/podbot-roadmap.md step 4.2.2
• ExecPlan: docs/execplans/4-2-2-safe-host-mounted-workspaces.md
• Lody session (this revision): https://lody.ai/leynos/sessions/f70986de-9ad6-42dc-92da-8c191adb561b
• Lody session (original 2026-05-29 draft): https://lody.ai/leynos/sessions/34745dd5-67cd-4cff-9cf5-4fa735dac79b

[coderabbitai]

Important

Review skipped

Draft detected.

Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Organization UI

Review profile: ASSERTIVE

Plan: Pro Plus

Run ID: 5f33f796-1391-4ff8-9c07-e208e80e3588

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch 4-2-2-safe-host-mounted-workspaces

---

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[sourcery-ai]

Reviewer's Guide

Adds a detailed ExecPlan document for roadmap step 4.2.2 that designs and constrains the future implementation of safe host-mounted workspaces, including config schema changes, a hexagonal domain/adapter split, validation rules, testing strategy, and rollout steps, but makes no code or behaviour changes yet.

Sequence diagram for planning and applying a host-mounted workspace

sequenceDiagram
actor Operator
participant CLI
participant Library
participant Probe as DefaultHostPathProbe
participant Engine as EngineConnector
participant Docker as Bollard

Operator->>CLI: Configure workspace.source=HostMount
CLI->>Library: plan_host_mount_workspace(AppConfig, Probe)
Library->>Probe: canonicalize(host_path)
Probe-->>Library: canonical_host_path
Library->>Probe: probe_writable(canonical_host_path)
Probe-->>Library: Result
Library-->>CLI: HostMountPlan
CLI->>Engine: create_container_async(HostMountPlan)
Engine->>Docker: create_container(HostConfig.mounts with Mount)
Docker-->>Engine: container_id
Engine-->>CLI: container_id
CLI-->>Operator: Container with host-mounted workspace

Loading

File-Level Changes

Change	Details	Files
Introduce an ExecPlan document that fully specifies how to implement safe host-mounted workspaces, including domain model, configuration, validation, engine integration, testing, and rollout process.	Describe current behaviour where host-mounted workspaces are configured but not actually mounted into containers, and define the desired end state with canonicalisation, allowlists, symlink-escape checks, and write-probe validation. Define architectural constraints and tolerances, including a hexagonal split with a HostPathProbe port, path-handling rules (cap-std, Utf8PathBuf), error-handling policies, and config-surface rules for host mounts. Specify new configuration surface ([mounts] section and PODBOT_MOUNTS_* env vars), validation semantics for allowlists and reserved container paths, and how planning should be exposed via plan_host_mount_workspace and HostMountPlan types. Lay out a milestone-based implementation plan covering config wiring, domain types and planner logic, default HostPathProbe adapter, engine integration with bollard::models::Mount, BDD/proptest/testcontainers coverage, documentation updates, and PR workflow. Capture risks, threat model, residual TOCTOU concerns, decision log entries (e.g., using HostConfig.mounts with typed Mount and RPRIVATE), and acceptance/quality criteria for when step 4.2.2 is considered complete.	docs/execplans/4-2-2-safe-host-mounted-workspaces.md

---

Tips and commands

Interacting with Sourcery

• Trigger a new review: Comment @sourcery-ai review on the pull request.
• Continue discussions: Reply directly to Sourcery's review comments.
• Generate a GitHub issue from a review comment: Ask Sourcery to create an
  issue from a review comment by replying to it. You can also reply to a
  review comment with @sourcery-ai issue to create an issue from it.
• Generate a pull request title: Write @sourcery-ai anywhere in the pull
  request title to generate a title at any time. You can also comment
  @sourcery-ai title on the pull request to (re-)generate the title at any time.
• Generate a pull request summary: Write @sourcery-ai summary anywhere in
  the pull request body to generate a PR summary at any time exactly where you
  want it. You can also comment @sourcery-ai summary on the pull request to
  (re-)generate the summary at any time.
• Generate reviewer's guide: Comment @sourcery-ai guide on the pull
  request to (re-)generate the reviewer's guide at any time.
• Resolve all Sourcery comments: Comment @sourcery-ai resolve on the
  pull request to resolve all Sourcery comments. Useful if you've already
  addressed all the comments and don't want to see them anymore.
• Dismiss all Sourcery reviews: Comment @sourcery-ai dismiss on the pull
  request to dismiss all existing Sourcery reviews. Especially useful if you
  want to start fresh with a new review - don't forget to comment
  @sourcery-ai review to trigger a new review!

Customizing Your Experience

Access your dashboard to:

• Enable or disable review features such as the Sourcery-generated pull request
  summary, the reviewer's guide, and others.
• Change the review language.
• Add, remove or edit custom review instructions.
• Adjust other review settings.

Getting Help

• Contact our support team for questions or feedback.
• Visit our documentation for detailed guides and information.
• Keep in touch with the Sourcery team by following us on X/Twitter, LinkedIn or GitHub.