Answer in brief
CVE-2026-54766 records a Medium severity vulnerability in Vikunja has a project duplication bypasses write-permission check on the target parent project. The current sources do not mark it as known exploited. The current feed maps code.vikunja.io/api (go). Check affected ranges and fixed versions before updating.
Analysis pending evidence review
HOL Guard separates source facts from reviewed analysis. See the methodology.
A CVSS score is not reported in the current record. The current sources do not mark it as known exploited. Treat this as a source-backed prioritization signal, not a statement about your environment.
Analysis status
Analysis pending evidence review
Factual feed record only; HOL analysis is not approved for indexing. Read the methodology.
The current feed maps code.vikunja.io/api (go). Check affected ranges and fixed versions before updating.
| Package | Affected range | Fixed version |
|---|---|---|
| code.vikunja.io/apigo | >=0.21.0,<=2.3.0 | 2.4.0 |
Published upstream
Aug 28, 2026
Evidence: source:ghsa:source_dates:source-dates:recordSource modified
Aug 28, 2026
Evidence: source:ghsa:source_dates:source-dates:recordFirst seen by HOL
Aug 28, 2026
## Summary The project-duplication endpoint fails to enforce write access to the target parent project. Any authenticated (non-link-share) user can duplicate a project they can read into **any** parent project on the instance, regardless of whether they have write access to that parent — injecting an attacker-owned project into another user's or team's project hierarchy. ## Details `ProjectDuplicate.CanCreate` (`pkg/models/project_duplicate.go`) is meant to require write access to the parent the duplicate is placed under — its own comment says "Parent project exists + user has write access". The implementation does neither correctly: ```go func (pd *ProjectDuplicate) CanCreate(s *xorm.Session, a web.Auth) (canCreate bool, err error) { pd.Project = &Project{ID: pd.ProjectID} canRead, _, err := pd.Project.CanRead(s, a) if err != nil || !canRead { return canRead, err } if pd.ParentProjectID == 0 { return canRead, err } // Parent project exists + user has write access to is (-> can create new projects) parent := &Project{ID: pd.ParentProjectID} return parent.CanCreate(s, a) // <-- bug } ``` Two defects compound here: 1. **Wrong permission method.** It calls `parent.CanCreate` ("may I create *this* project?") instead of `parent.CanWrite` ("may I create children *inside* this project?"). The latter is what the normal create path uses — `POST /projects` with a `parent_project_id` enforces `parent.CanWrite` via `Project.CanCreate` (`pkg/models/project_permissions.go:196-199`). 2. **Unhydrated struct.** `parent` is constructed as `&Project{ID: pd.ParentProjectID}` and never loaded from the database, so its in-memory `ParentProjectID` is always `0`. Inside `Project.CanCreate` the only branch that performs any permission check is `if p.ParentProjectID != 0 { return parent.CanWrite(...) }` — which therefore never executes. Control falls through to the link-share check and then `return true, nil`. The result is `true` for any authenticated non-link-share user, for any `ParentProjectID`. Nothing downstream re-checks: `ProjectDuplicate.Create` → `CreateProject` → `checkProjectBeforeUpdateOrDelete` (`pkg/models/project.go:954`) validates only that the parent exists, is not a pseudo-project, and introduces no cycle — no authorization. ## Impact An authenticated user can: - Duplicate any project they can read (including their own) and attach the copy as a child of **any** parent project ID on the instance, with no write access to that parent. - Inject an attacker-owned project into other users'/teams' project trees. The duplicate is owned by the attacker but appears inside the victim's hierarchy; members of the victim parent see it, and because Vikunja propagates parent access down the tree, they may inherit access to the injected project — enabling content injection / spam / phishing inside another tenant's workspace. This is a bypass of the same parent-write guard that the ordinary create path enforces, so the duplicate route is an authorization hole for an operation that is otherwise correctly gated. The endpoint requires authentication; it does not expose or modify the victim's existing project data (the source is attacker-readable), so the impact is an integrity / access-control violation rather than confidentiality. ## Proof of Concept 1. As user A, create or have read access to any project `S` (e.g. id 100). 2. Identify a parent project `P` (e.g. id 5) owned by user B, to which A has **no** access. 3. Call `PUT /api/v1/projects/100/duplicate` with body `{"parent_project_id": 5}`. 4. The request succeeds (201). A new project owned by A is created as a child of B's project 5, despite A having no write access to it. The equivalent `POST /api/v1/projects` with `parent_project_id: 5` would be correctly rejected with 403. ## Affected versions Introduced with the namespace→project migration (commit `fef253312`, first released in v0.21.0) and present through the latest release (v2.3.0). The shared model also backs the new `/api/v2` duplication route under review, so any v2 release would inherit the same flaw unless fixed in the model. ## Recommended Fix In `ProjectDuplicate.CanCreate`, check write access to the parent directly: ```go parent := &Project{ID: pd.ParentProjectID} return parent.CanWrite(s, a) ``` `Project.CanWrite` loads the project from the database and evaluates real permissions, fixing both the wrong-method and the unhydrated-struct defects at once and matching the documented contract. (It also rejects archived parents, which is desirable.)
Quoted source text, attributed separately from HOL analysis.