gemini extensions install error extension failed to load context missing fails

Not specified by the original author.

TL;DR

For "loadSourceAsModule - Failed to load file "C:/.../Extensions-32448/OpenDose3D/lib/Slicer-5.6/qt-scripted-modules/OpenDose3D.py" as module "Op": restart Slicer, then check the Extensions Manager's Installed tab. (If you installed from a file per section 1 and the module still doesn't appear after a restart, you are in section 2 territory, verify the package revision.) In Slicer's Extensions Manager, use Install extension from file (wrench/toolbox menu) and select the downloaded package. OpenDose3D packages that predate a Slicer 5.8 update and no longer load under it.

loadSourceAsModule - Failed to load file "C:/.../Extensions-32448/OpenDose3D/lib/Slicer-5.6/qt-scripted-modules/OpenDose3D.py" as module "OpenDose3D" !
Fail to instantiate module "OpenDose3D"
The following modules failed to be instantiated: Calibration OpenDose3D Gamma

When to use

You are seeing this: Extensions are built for a specific Slicer revision (e.g. Use this skill when you run into "Diagnose extension installation failures in 3D Slicer".

When not to use

If your error message or symptom does not match what is described above, this is probably not your fix. Search for your exact error text instead of forcing this one to fit.

Versions

Versions mentioned in the source: Slicer 5.4.0, Slicer 5.6.2. If you are on something much newer or older, the details may have shifted.

Use this when an extension won't download, installs but its modules never appear, or the Extensions Manager itself misbehaves. Extensions are built for a specific Slicer revision (e.g. Extensions-31938 under Slicer 5.4.0, Extensions-32448 under Slicer 5.6.2), most failures are a mismatch between the extension package and your Slicer build, or a network problem. Find your revision in Help / About; the catalog uses it in the URL (extensions.slicer.org/catalog/All/32448/win).

1. Failed downloading: https://slicer-packages.kitware.com/...

The Extensions Manager could not fetch the package, this is a network problem, not a broken extension. It shows up on locked-down machines, proxies, and cloud VMs (firewalled govcloud instances hit it for every extension). Fix:

1. Download the extension package from the catalog website on a machine that has network access. Pick the entry matching your OS and your Slicer revision.

2. In Slicer's Extensions Manager, use Install extension from file (wrench/toolbox menu) and select the downloaded package.

3. Restart Slicer.

The same offline path works on air-gapped machines; the documented section is "Install extensions without network connection" in the user guide.

2. Installed, but the module never appears / loadSourceAsModule - Failed to load

Real symptom, from the Error Log:

loadSourceAsModule - Failed to load file "C:/.../Extensions-32448/OpenDose3D/lib/Slicer-5.6/qt-scripted-modules/OpenDose3D.py" as module "OpenDose3D" !
Fail to instantiate module "OpenDose3D"
The following modules failed to be instantiated: Calibration OpenDose3D Gamma

This means the extension package was built against a different Slicer version than the one running, e.g. OpenDose3D packages that predate a Slicer 5.8 update and no longer load under it. Check, in order:

1. Does your Slicer's revision match the revision the package was built for? Extension folders are named Extensions-[revision]; a folder whose revision differs from Help / About is the wrong package.

2. If the versions genuinely differ, either install the Slicer version the extension targets, or check the extension's own issue tracker, extension authors announce compatibility updates there.

Note the distinction from the next case: here the load is attempted and fails. If no load is attempted at all, see below.

3. Extension shows "installed" but nothing new in the module list

Extensions are not activated until Slicer restarts. This is the most common "it installed but I can't find it" report: the install completed, the module only registers on the next launch. Restart Slicer, then check the Extensions Manager's Installed tab. (If you installed from a file per section 1 and the module still doesn't appear after a restart, you are in section 2 territory verify the package revision.)

4. Extensions Manager crashes or shows a blank/no-connection page

On Linux, opening the Extensions Manager has crashed Slicer outright (QtWebEngine + OpenGL stack traces ending in QQuickWidget/RenderWidgetHostViewQtDelegateWidget segfaults, after [Qt] An OpenGL Core Profile was requested, but it is not supported warnings). This is a rendering-environment problem, not a catalog problem:

• Update the graphics driver / enable a supported OpenGL profile if you can.

• Otherwise skip the in-app browser entirely: use the offline install-from-file flow from section 1.

5. Extension missing from the catalog for your revision

Some extensions are only built for certain platforms or for preview (nightly) builds. Search the catalog for your exact revision and OS; if the extension is absent there but exists for the nightly, install a Slicer preview release and retry. If it is absent for your OS entirely, the extension does not support that platform (check its repository for a platform-support statement before assuming the catalog is broken).

Quick triage checklist

1. Capture the exact symptom: download error, load error in the Error Log, missing-after-restart, or manager crash.

2. Network → offline install from file (section 1).
3. loadSourceAsModule failure → revision mismatch (section 2).
4. Installed-but-missing → restart Slicer (section 3).
5. Manager crash → graphics environment; use offline install (section 4).
6. Not in catalog → platform/nightly coverage (section 5).

Built from 3D Slicer Discourse threads: 'Extension download issues in govcloud' (Failed downloading + offline install-from-file procedure), 'OpenDose3D module is not loaded' (loadSourceAsModule / failed-to-instantiate on a version mismatch, Extensions-32448 under Slicer 5.6.2), 'Opening extension manager makes Slicer crash on Linux' (QtWebEngine OpenGL crash), and the Extensions Manager user guide's 'Install extensions without network connection' section.

Not specified by the original author.

TL;DR

The extension tool and parameter names collided with something Gemini CLI rejected. Miro updated their tool and parameter names to fix it: restart Gemini CLI so it re-fetches the tool list from the extension, and chat works again.

Fix

1. The extension tool and parameter names collided with something Gemini CLI rejected. Expected: You get the expected result; the problem is gone.

2. Miro updated their tool and parameter names to fix it: restart Gemini CLI so it re-fetches the tool list from the extension, and chat works again. Expected: You get the expected result; the problem is gone.

3. If the 400 persists, make sure you are on the latest version of the miro-ai extension. Expected: You get the expected result; the problem is gone.

When to use

• You are dealing with this situation.
• The symptom matches: Miro extension breaks Gemini CLI chat with 400 INVALID_ARGUMENT.

When NOT to use

• Unrelated issues with a different cause.
• You need general documentation for the tool; check the official docs instead.

Compatibility

Reported per the linked source.

Variant phrasings

Miro extension breaks Gemini CLI chat with 400 INVALID_ARGUMENT

Why it happens

GitHub issue on miroapp/miro-ai: after installing the miro-ai extension, every Gemini CLI chat attempt fails with a 400 INVALID_ARGUMENT API error until the extension is removed.

Edge cases

• If your error message differs even slightly, this is probably a different issue; search the exact text.
• If the fix does not help, capture the full error output and check the source link for updates.

GitHub issue on miroapp/miro-ai: after installing the miro-ai extension, every Gemini CLI chat attempt fails with a 400 INVALID_ARGUMENT API error until the extension is removed. Resolved: The extension tool and parameter names collided with something Gemini CLI rejected. Miro updated their tool and parameter names to fix it: restart Gemini CLI so it re-fetches the tool list from the extension, and chat works again. If the 400 persists, make sure you are on the latest version of the miro-ai extension. Source: https://github.com/miroapp/miro-ai/issues/10

Not specified by the original author.

TL;DR

Lab 4 moved to a new extension API, so extensions built for Lab 3 declare an incompatible peer range and the installer refuses them. Find the extension release whose compatibility range includes your Lab version and install that exact version. If no Lab 4 release exists, either stay on Lab 3 for that extension or drop the extension; forcing the install produces a broken frontend.

jupyter labextension install: version conflict with jupyterlab 4

Use this when

• jupyter labextension install [name] fails with a peer-dependency or version-range conflict naming jupyterlab 4.
• The extension's release notes mention Lab 3 support but nothing about Lab 4.
• You upgraded Lab and a previously working extension now refuses to install.

Not for this skill when

• The install succeeds but the extension does not appear in Lab. That is a build or load problem, not a version conflict.
• The build runs out of memory. That is the webpack OOM case, a different fix.

Steps

1. Record exact versions: jupyter lab --version and the extension version you tried to install. Expected: you know the Lab 4.x line and the extension's version.
2. Check the extension's release history for a Lab 4-compatible release (its changelog or compatibility matrix names the supported Lab range). Expected: you find the first extension version that supports Lab 4.
3. Install that version explicitly: jupyter labextension install [name]@[lab4-compatible-version]. Expected: the installer resolves without a conflict error.
4. Rebuild Lab: jupyter lab build. Expected: the build completes and lists the extension among the included packages.
5. Verify in the running Lab: open the extension manager or check jupyter labextension list. Expected: the extension shows as enabled and valid for the running Lab version.
6. If no Lab 4 release exists, decide explicitly: pin Lab to 3.x while you need the extension, or remove the extension and move to Lab 4. Expected: a deliberate choice recorded in the environment spec, not a forced install.

Tool compatibility

JupyterLab 4.x (and the 3.x line for the downgrade path); labextension installer as shipped with Lab.

Variant phrasings

labextension incompatible with jupyterlab 4

The general phrasing. Install the extension release built for the Lab 4 API.

jupyterlab 4 extension version conflict

The conflict phrasing. The fix is version selection, never forcing.

extension requires jupyterlab 3 but have 4

The peer-range phrasing. Either downgrade Lab or find the newer extension release.

Why it happens

JupyterLab extensions declare the Lab versions they work with, and Lab 4 introduced breaking changes in the extension API (newer Lumino, new token contracts), so Lab 3 extensions honestly report incompatibility. The installer's refusal is the system working as designed: it prevents a frontend that would fail to load at runtime. The correct move is version alignment in either direction. Forcing past the check with flags or manual package edits bypasses the only guardrail and yields a Lab that builds but shows a broken or missing extension, which is harder to diagnose than the original clean error.

Edge cases / pitfalls

• Prebuilt (pip/conda) extensions and source (npm) extensions version independently. Make sure the Lab 4-compatible release exists in the channel you install from.
• Pinning Lab to 3.x for one extension holds back security and feature updates. Revisit the pin on a schedule.
• After switching extension versions, do a clean rebuild. Stale build artifacts from the old version can linger and confuse verification.
• Agents that install extensions in Dockerfiles should pin both the Lab version and the extension version, or the next rebuild can drift into a conflict.

Provenance

Resolved from the public thread: https://vectle.com/posts/pst_sg0FHibY2eNrnU2Rp5WQMg

Resolved from the public thread: https://vectle.com/posts/pst_sg0FHibY2eNrnU2Rp5WQMg

Thread query: jupyter labextension install: version conflict with jupyterlab 4

TL;DR: Lab 4 moved to a new extension API, so extensions built for Lab 3 declare an incompatible peer range and the installer refuses them. Find the extension release whose compatibility range includes your Lab version and install that exact version. If no Lab 4 release exists, either stay on Lab 3 for that extension or drop the extension; forcing the install produces a broken frontend.