ComfyUI-Easy-Use: unauthenticated path traversal (CVE-2026-123790)
An unauthenticated caller can abuse the affected ComfyUI-Easy-Use preview-save handler to overwrite a custom node and potentially execute code. Update the extension to protect the files ComfyUI loads into its process.
Public CVE record · Native module source
| Field | Value |
|---|---|
| Affected component and versions | ComfyUI-Easy-Use · 1.3.6 through 1.4.1 |
| Fixed release | See the advisory |
| Evidence status | Fixture comparison recorded, CVE coverage unverified |
| Test environment | Lab fixture or affected build documented in the evidence below |
| Eden release and module version | 1.103.0 |
| Tested at | 2026-11-12 UTC |
| First publicly published | 2026-11-12T17:20:32Z |
cve.now entry. The experiment used a synthetic HTTP fixture.
Eden terminates the HTTP connection to the affected product, evaluates a declarative policy and — in tier 2 — a separately compiled native Rust module under signed module authority, before any byte reaches the application. It does not patch the affected component; the vendor's fixed release remains the durable repair.
Status: Two-tier lab fixture comparison recorded. The fixture ran against Eden 1.103.0.
The vulnerability
The advisory describes an unauthenticated traversal in the preview-save handler that can overwrite a custom node. The synthetic fixture recorded a file-read marker when a node body carried traversal fragments.
The Eden patch
Tier 1 — declarative policy (blunt, fail-closed): deny every request under
/easy-use (non-slash scope), legitimate traffic included — false positives
measured below.
Tier 2 — native Rust module gabriel_comfytrav_v1 (surgical):
refuses traversal fragments in path parameters
Implement the remediation
Use this procedure in an isolated fixture to reproduce the scoped restriction. Follow the vendor update for the underlying exposure. See deployment prerequisites.
- Prepare the target. Verify the loaded product and version against the advisory. Use a compatible Eden HTTP interlay and an authenticated operator with Configure permission. Use this reconstruction of the report’s scoped fixture deny.
- Apply the declarative policy. Scope the policy to
/easy-useand descendants (all methods,deny). SetINTERLAYto the target interlay UUID in the authenticated Eden CLI. Run an authenticatedGET /api/v1/interlays/{interlay}withX-Eden-Verbose: true. Save its complete current settings object ascurrent-settings.jsonand retain the prior remediation declaration for recovery. SetIDEMPOTENCY_KEYto a fresh key for this intended mutation.
Save the declaration as remediation.json.
{
"version": 1,
"profile": "http-structure-v1",
"policy": {
"version": 1,
"rules": [
{
"scope": {
"methods": [],
"path": "/easy-use",
"descendants": true
},
"require": {
"kind": "deny"
}
}
]
}
}
jq -e --slurpfile policy remediation.json \
'.remediation = $policy[0] | {settings: .}' current-settings.json \
> interlay-update.json
eden api request PATCH "/api/v1/interlays/$INTERLAY" \
--header "Idempotency-Key:$IDEMPOTENCY_KEY" \
--body-file interlay-update.json --content-type application/json --confirm
- Verify the serving policy. Read back the interlay.
eden api request GET "/api/v1/interlays/$INTERLAY" \
--header 'X-Eden-Verbose:true'
Check that remediation_status.state is ready and both desired and effective SHA-256 values match the intended declaration. Replay the benign and marked requests in the recorded comparison on an isolated target. Check the recorded response and backend outcome for each probe.
4. Restore the previous policy. Coordinate the change with any other operator updating this interlay. Read current settings again and merge the saved remediation field into them. If the prior settings had no remediation field, omit it in the replacement. Send the same PATCH with a fresh idempotency key, then read back the intended state. Repeat the verification probes. Keep the affected route isolated if restoring the prior settings reopens exposure.
Vendor update. Use the advisory to identify the supported update or workaround for the installed release. The advisory data used here supplies no fixed version. Confirm the running version and verify the affected feature after the change.
Measured comparison (2026-11-12 UTC)
Summary of measured outcomes across the full authority lifecycle (baseline → tier-1 → tier-2 → revoke → restore):
- Baseline (no remediation): benign 200, attack 200 with the primitive recorded at the fixture oracle
- Tier 1 (deny): both probes refused 403 (blunt, false positives included)
- Tier 2 (module): benign 200 / attack 403, zero backend effects
Authority lifecycle and negatives
Import 201, activation: explicit_required; publisher enroll_key, host
inventory + approve, reader inspect refused 403; tampered-statement admit
400; invalid-signature admit 403; admit prepared (not serving), activate
→ receipt serving_now, all expected shards acknowledged the operation. Revoke → the identical attack
gets 503 (fail-closed); re-approve + re-activate → 403 again.
Test scope
The recorded results apply to the simulated requests and the fixture's backend markers. Coverage of the vendor's vulnerable code path remains unverified.
Performance
The public evidence for this case does not include request-latency measurements.