GHSA-PXCC-8665-PHX8

Vulnerability from github – Published: 2026-06-26 22:29 – Updated: 2026-06-26 22:29
VLAI
Summary
YARD static cache reads raw traversal paths before router sanitization
Details

Summary

YARD's static cache lookup reads a request path before the router's path cleanup runs. When a server is configured with a document root, a traversal path such as /../yard-cache-secret.html is joined against that root and can return a readable sibling .html file outside the intended static tree.

The potential security risk seems low, as only html-ending files can be read, but still the risk of reading arbitrary html files is a confiendtiality issue in itself, which is why we decided to report. Please let us know if this is out of your project's scope.

Details

The --docroot CLI option stores the configured directory in server_options[:DocumentRoot] at lib/yard/cli/server.rb:198, and adapter initialization copies that value into adapter.document_root at lib/yard/server/adapter.rb:76. For Rack requests, RackAdapter#call builds a request object from the Rack environment at lib/yard/server/rack_adapter.rb:58 and passes it to router.call(request) at lib/yard/server/rack_adapter.rb:60. Router#call then stores the incoming request at lib/yard/server/router.rb:55 and invokes check_static_cache before normal routing at lib/yard/server/router.rb:56. Inside check_static_cache, the only initial guard is that adapter.document_root is present at lib/yard/server/static_caching.rb:35; the cache path is built from File.join(adapter.document_root, request.path.sub(/\.html$/, '') + '.html') at lib/yard/server/static_caching.rb:36, without cleaning .. components first. If that resolved path is a regular file, File.file? accepts it at lib/yard/server/static_caching.rb:38 and the file bytes are returned as a 200 HTML response at lib/yard/server/static_caching.rb:40. The later route sanitizer in final_options uses File.cleanpath(...).gsub(...) at lib/yard/server/router.rb:181 and lib/yard/server/router.rb:182, but a static-cache hit returns before that code is reached.

PoC

poc.zip

bash ./poc/run.sh

expected output:

run 1: exit=0 timed_out=False duration=0.08s matched=True phase=oracle fingerprint='YARD_STATIC_CACHE_PATH_TRAVERSAL'
run 2: exit=0 timed_out=False duration=0.08s matched=True phase=oracle fingerprint='YARD_STATIC_CACHE_PATH_TRAVERSAL'
run 3: exit=0 timed_out=False duration=0.08s matched=True phase=oracle fingerprint='YARD_STATIC_CACHE_PATH_TRAVERSAL'

The YARD_STATIC_CACHE_PATH_TRAVERSAL fingerprint is emitted only after the PoC observes a 200 static-cache response whose body contains the sibling file outside the configured document root. A setup failure, syntax failure, or cache miss would not print this oracle and would not demonstrate this traversal read.

Impact

A remote unauthenticated HTTP client who can reach a YARD documentation server with DocumentRoot/--docroot enabled can request .html paths containing parent-directory components and receive readable matching files outside the configured document root. The required guards are narrow: adapter.document_root must be set, the traversed target must exist as a regular readable file, and the target must be reachable through the implementation's forced .html suffix. Those requests bypass the later final_options path cleanup because the cache check runs first. The resulting severity class is information disclosure: response bodies can contain off-root .html file contents, but this path does not show write access, code execution, or arbitrary files without the .html constraint.

Show details on source website

{
  "affected": [
    {
      "package": {
        "ecosystem": "RubyGems",
        "name": "yard"
      },
      "ranges": [
        {
          "events": [
            {
              "introduced": "0"
            },
            {
              "fixed": "0.9.44"
            }
          ],
          "type": "ECOSYSTEM"
        }
      ]
    }
  ],
  "aliases": [
    "CVE-2026-49342"
  ],
  "database_specific": {
    "cwe_ids": [
      "CWE-22"
    ],
    "github_reviewed": true,
    "github_reviewed_at": "2026-06-26T22:29:29Z",
    "nvd_published_at": "2026-06-19T20:16:18Z",
    "severity": "MODERATE"
  },
  "details": "### Summary\nYARD\u0027s static cache lookup reads a request path before the router\u0027s path cleanup runs. When a server is configured with a document root, a traversal path such as `/../yard-cache-secret.html` is joined against that root and can return a readable sibling `.html` file outside the intended static tree. \n\nThe potential security risk seems low, as only html-ending files can be read, but still the risk of reading arbitrary html files is a confiendtiality issue in itself, which is why we decided to report. Please let us know if this is out of your project\u0027s scope.\n\n### Details\nThe `--docroot` CLI option stores the configured directory in `server_options[:DocumentRoot]` at `lib/yard/cli/server.rb:198`, and adapter initialization copies that value into `adapter.document_root` at `lib/yard/server/adapter.rb:76`. For Rack requests, `RackAdapter#call` builds a request object from the Rack environment at `lib/yard/server/rack_adapter.rb:58` and passes it to `router.call(request)` at `lib/yard/server/rack_adapter.rb:60`. `Router#call` then stores the incoming request at `lib/yard/server/router.rb:55` and invokes `check_static_cache` before normal routing at `lib/yard/server/router.rb:56`. Inside `check_static_cache`, the only initial guard is that `adapter.document_root` is present at `lib/yard/server/static_caching.rb:35`; the cache path is built from `File.join(adapter.document_root, request.path.sub(/\\.html$/, \u0027\u0027) + \u0027.html\u0027)` at `lib/yard/server/static_caching.rb:36`, without cleaning `..` components first. If that resolved path is a regular file, `File.file?` accepts it at `lib/yard/server/static_caching.rb:38` and the file bytes are returned as a `200` HTML response at `lib/yard/server/static_caching.rb:40`. The later route sanitizer in `final_options` uses `File.cleanpath(...).gsub(...)` at `lib/yard/server/router.rb:181` and `lib/yard/server/router.rb:182`, but a static-cache hit returns before that code is reached.\n\n### PoC\n[poc.zip](https://github.com/user-attachments/files/28185421/poc.zip)\n\n```bash\nbash ./poc/run.sh\n```\nexpected output:\n```text\nrun 1: exit=0 timed_out=False duration=0.08s matched=True phase=oracle fingerprint=\u0027YARD_STATIC_CACHE_PATH_TRAVERSAL\u0027\nrun 2: exit=0 timed_out=False duration=0.08s matched=True phase=oracle fingerprint=\u0027YARD_STATIC_CACHE_PATH_TRAVERSAL\u0027\nrun 3: exit=0 timed_out=False duration=0.08s matched=True phase=oracle fingerprint=\u0027YARD_STATIC_CACHE_PATH_TRAVERSAL\u0027\n```\n\nThe `YARD_STATIC_CACHE_PATH_TRAVERSAL` fingerprint is emitted only after the PoC observes a `200` static-cache response whose body contains the sibling file outside the configured document root. A setup failure, syntax failure, or cache miss would not print this oracle and would not demonstrate this traversal read.\n\n### Impact\nA remote unauthenticated HTTP client who can reach a YARD documentation server with `DocumentRoot`/`--docroot` enabled can request `.html` paths containing parent-directory components and receive readable matching files outside the configured document root. The required guards are narrow: `adapter.document_root` must be set, the traversed target must exist as a regular readable file, and the target must be reachable through the implementation\u0027s forced `.html` suffix. Those requests bypass the later `final_options` path cleanup because the cache check runs first. The resulting severity class is information disclosure: response bodies can contain off-root `.html` file contents, but this path does not show write access, code execution, or arbitrary files without the `.html` constraint.",
  "id": "GHSA-pxcc-8665-phx8",
  "modified": "2026-06-26T22:29:29Z",
  "published": "2026-06-26T22:29:29Z",
  "references": [
    {
      "type": "WEB",
      "url": "https://github.com/lsegal/yard/security/advisories/GHSA-pxcc-8665-phx8"
    },
    {
      "type": "ADVISORY",
      "url": "https://nvd.nist.gov/vuln/detail/CVE-2026-49342"
    },
    {
      "type": "WEB",
      "url": "https://github.com/lsegal/yard/commit/f78c19f0dd33a407085b4ed181bb60c0aa0078b4"
    },
    {
      "type": "PACKAGE",
      "url": "https://github.com/lsegal/yard"
    }
  ],
  "schema_version": "1.4.0",
  "severity": [
    {
      "score": "CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:L/I:N/A:N",
      "type": "CVSS_V3"
    }
  ],
  "summary": "YARD static cache reads raw traversal paths before router sanitization"
}



Log in or create an account to share your comment.




Tags
Taxonomy of the tags.


Loading…

Loading…

Loading…

Drag nodes to arrange them. Double-click a related vulnerability to open it, or a source to visit its link. The Physics button on the left tunes the layout and the simulation; the legend at the bottom left hides or shows nodes by kind and by sighting type.

Forecast uses a logistic model when the trend is rising, or an exponential decay model when the trend is falling. Fitted via linearized least squares.

Sightings

Author Source Type Date Other

Nomenclature

  • Seen: The vulnerability was mentioned, discussed, or observed by the user.
  • Confirmed: The vulnerability has been validated from an analyst's perspective.
  • Published Proof of Concept: A public proof of concept is available for this vulnerability.
  • Exploited: The vulnerability was observed as exploited by the user who reported the sighting.
  • Patched: The vulnerability was observed as successfully patched by the user who reported the sighting.
  • Not exploited: The vulnerability was not observed as exploited by the user who reported the sighting.
  • Not confirmed: The user expressed doubt about the validity of the vulnerability.
  • Not patched: The vulnerability was not observed as successfully patched by the user who reported the sighting.
Vulnerability timeline

Sightings and recorded dates, grouped by day · UTC

JSON
How durations are calculated

Following the VTEM framework (Black Hat USA 2015), time to patch starts when the vendor or a coordinator was informed, total time to patch at discovery, and disclosure to patch at public disclosure; vendor response time runs from the report to its acknowledgement. Time to exploit and time unpatched start at the earliest discovery or report, falling back to public disclosure. Unpatched time ends at the first recorded patch, or the calculation time when no patch is recorded. Exploitation to patch shows the interval between those two events; a published proof of concept is a public exploit, not exploitation.

Each result shows its source dates. A negative interval means the end event came first; when date-only inputs leave the order open, the result says so. Date-only inputs are approximate. The earliest evidence wins: KEV report dates and sightings are labelled as such. Records describing another vulnerability are listed but never calculated from, and a PoC alone does not establish exploitation. Missing patch evidence does not prove that no patch exists. Filters below change the event list only.

Events included in the timeline

Loading timeline…


    Loading…

    Loading…

    Loading…

    Related by attack behaviour

    Vulnerabilities whose description is nearest to this one in the vector space of the CIRCL/vulnerability-attack-technique-biencoder model. This is a similarity search over the bi-encoder space (plain cosine), not a classification, and it has no measured accuracy.