[DMontgomery40]

David's AirPods Max desk hanger printed upright with no supports. The design pass planned it on one 80 mm side, but Counterpart's print step ran lay_flat, which put it back on its desk-side bottom. It then sliced with the stock 0.20mm Standard preset (enable_support 0), so the 40 mm top jaw printed in mid-air. David: "WE ARE JUST A BRIDGE. JUST SEND THE THING TO SLICER, LET IT SLICE IT, PRINT WHAT THE SLICER PUKED OUT."

Now Counterpart sends the model to the slicer as it is. Following David's choices, the slicer orients it and the designer chooses supports:

• No more lay_flat. Every STL goes to slice_stl with orient: true, so Bambu Studio's own auto-orient places it. A 3MF keeps its placement, as before.
• Supports come from the designer. A design pass answers supports (enabled, plus tree(auto) or normal(auto)). Counterpart stores it with the model and sends it as enable_support / support_type, including when supports are off. A model with no choice (scans, older designs, published models) leaves the printer profile in charge, as before.
• Other slicers get the model as it is. orient and the support arguments go only to Bambu-compatible slicing. mcp-3D-printer-server ignores orient on PrusaSlicer and Cura and refuses support arguments there, so a supports choice on that route gets the same refusal.
• An MCP without those arguments is refused before slicing, with the existing print-settings-unsupported refusal: "Update bambu-printer-mcp in Settings → Printers so Counterpart can send this design's supports to the slicer." The support arguments are part of the slice-cache key and the estimate key.

Depends on:

• bambu-printer-mcp 1.1.54 (#101) and mcp-3D-printer-server 1.2.28 (#55) add the slice_stl support arguments. Both are now npm latest. The MCP owner checked by slicing only (Bambu Studio 02.08.02.61, P1S, no printer) that with orient: true the override reaches each object's own settings and the toolpaths: on a 3D cross, supports on gives 127 Support features and off gives none.
• The Mac runs bambu-printer-mcp 1.1.52, so it needs 1.1.54 installed before designs with supports can print.
• Merge order: this PR goes next, ahead of fix: keep a multi-colour design's 3MF as the model it prints #293, after fix: a print that never started ends its Live Activity; the old route bug's workarounds are gone #302. Rebased onto main d9cd9ee (fix: a print that never started ends its Live Activity; the old route bug's workarounds are gone #302), keeping fix: a print that never started ends its Live Activity; the old route bug's workarounds are gone #302's slicing progress requests and its HMS 0007 and connectionMode removals. fix: keep a multi-colour design's 3MF as the model it prints #293 rebases onto this PR. Its multi-colour 3MF keeps its placement, so the matching worker-rule sentence belongs with fix: keep a multi-colour design's 3MF as the model it prints #293 or a follow-up.

#298 port (first commit, d098003). #298 merged into feat/testflight-readiness, not main. This PR brings its runtime guidance to main for invocation.ts, touchup.ts, mcp-client.ts, bound-fusion.ts and their tests only; provider-runner and project-worker changes stay on the release branch. The hunk that overlaps #298 when the branches are combined is one rule line in invocation.ts: #298's "This pass exports one STL … re-orients it with lay_flat … not kept yet" becomes the bridge wording. The supports schema field is a separate block after ask.

Overlap with the release branch's provider commit c59d83b (kept on release by df5facec): in runner.ts, that commit changes interface ResultEvent to export interface, the ProjectWorker type and the inv.worker call. This PR doesn't edit those lines. Its runner.ts hunks are the TouchupSupports type and TouchupResult.supports near the top, and the supports parse at the start of structured().

Known risk of auto-orient, recorded for David:

• On the stove gap filler, slice_stl with orient: true failed with Studio's "empty layer" error. That is why lay_flat stayed primary in KAN-223. Such a model is now refused with Studio's own words before anything reaches the printer, rather than printed.
• Research this session found that Printago, MakerWorld cloud slicing and 3DPrinterOS keep the designed or uploaded orientation, and that Studio's --orient minimises overhang area without checking cantilevers. If David prefers the designer's pose, it's a one-line change to orient plus the worker rule.

Checks on 8c3ec95, with Codex clean on that exact head:

• printer-handoff: 116 pass, 0 fail, 5 skipped (fix: a print that never started ends its Live Activity; the old route bug's workarounds are gone #302 removed one test). This includes a test that a kept slice is reused only for the same supports choice.
• Companion: the same 18 tests fail on d098003 (main plus the Design passes plan the print with the printer MCP's FDM guidance #298 port) without this change (Pillow, USD, Blender repair, nearby pairing), and nothing new fails.
• The opt-in real mcp-3D-printer-server tests fail the same way before and after this change against the stale local checkout. The changed-only repo gate runs on the merge owner's Linux host. No printer, Mac UI or install was touched. Acceptance is David pressing Print in the installed app.

Created with Claude Opus 5.5 in Claude Code.

🤖 Generated with Claude Code

[chatgpt-codex-connector]

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review	Status	Commit	Review trigger
📝 Code Review	✅ Completed 2026-11-04T12:33:52.851170Z	8c3ec95	Manual request

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

[DMontgomery40]

@codex review

[chatgpt-codex-connector]

Codex Review: Didn't find any major issues. Keep it up!

Reviewed commit: bebaa791ae

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[cursor]

Reviewed SHA: bebaa791aefb550c1799275976416f08af56ad14 (matches PR head at review time)

Verdict: Ready

Findings: None

Validation: Reviewed git diff origin/main...bebaa791aefb550c1799275976416f08af56ad14 (20 files). Ran npm ci && npm test in services/printer-handoff (115 passed, 5 skipped). Ran npm run build in services/companion and node --test tests/objects-print.test.mjs (23 passed, including the new supports/orient bridge test). scripts/ci_local.py reports printer-handoff/printer-handoff node 22 PASS for this tree.

[DMontgomery40]

@codex review

[chatgpt-codex-connector]

Codex Review: Didn't find any major issues. What shall we delve into next?

Reviewed commit: 2f6bfd810c

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[DMontgomery40]

@codex review

[chatgpt-codex-connector]

Codex Review: Didn't find any major issues. Already looking forward to the next diff.

Reviewed commit: 353f3c24bb

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[DMontgomery40]

@codex review

[chatgpt-codex-connector]

Codex Review: Didn't find any major issues. Already looking forward to the next diff.

Reviewed commit: 54e7b83f5f

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[DMontgomery40]

@codex review

[chatgpt-codex-connector]

Codex Review: Didn't find any major issues. Keep it up!

Reviewed commit: 8c3ec95a73

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[cursor]

Reviewed SHA: 8c3ec95a73fab883db3855cdf9c3c53e033fba85 (matches PR head)

Verdict: Ready

Findings: None.

Validation: Read diff main...8c3ec95; traced companion designedSupports / touchup persistence, printer-handoff slice path (removed lay_flat, orient: true, support args, slice-cache key), and bound-Fusion printer-tool proxy. Built and ran services/printer-handoff tests (bound-fusion.test.mjs, printer-handoff.test.mjs — 81/81 pass). Built companion packages and ran objects-print.test.mjs subtest "the designer's supports choice reaches the slicer as it is…" (pass). Attempted broader companion touchup tests; failures were environment PIL missing, not this diff.