Tomcat partial-PUT path traversal

Tomcat partial-PUT path traversal (CVE-2025-24813) through Eden's API

Source and API bundle

Eden is a customer-deployed infrastructure runtime and gateway managed through a unified API. Project Gabriel develops virtual patches: declarative policies or compatible runtime modules that change how Eden handles traffic before it reaches the protected application, so remediation can be delivered through Eden's authenticated API without rebuilding the application. This case examines removing the write path that lets a caller place a file with a maliciously named partial PUT and have it deserialized by a Tomcat session handler.

Remediated 2026-11-11 through Eden's API. We deployed the http-structure-v1 deny rule scoped to PUT under /uploads via the authenticated interlay PATCH API, confirmed desired == effective, state: ready, and verified with the attack delivery that the malicious partial PUT is refused at the boundary (403) with zero deserialization effects, while static content and ordinary POST uploads on the same deployment still work. RBAC and revoke causality verified. This removes the file-placement prerequisite; the deserialization defect itself is repaired only by the patched Tomcat, stated below. Assessment: 10 November 2026 UTC.

The vulnerability

In affected Tomcat configurations, the combination of write-enabled default servlets, partial PUT support and a serialized-session file store lets a caller write attacker-controlled bytes to a path whose filename is then used as a deserialization session id, achieving remote code execution when classes exist on the classpath (primary advisory). The prerequisites are a specific method (PUT) on specific paths plus the session-file store. The durable fix is a patched Tomcat release or removing the configuration. Eden can refuse the write path before it reaches the backend, which removes the file-placement prerequisite for the paths that traverse it.

The challenge of patching the fleet

Tomcat embeds in many products, and the vulnerable configuration (default servlet write + partial PUT + serialized sessions) appears in deployments the application owner cannot easily rebuild. Upgrading means coordinating library compatibility, web.xml differences and rolling restarts per installation. Removing the write capability outright disables legitimate uploads in some deployments, so a blanket disable is a functional restriction that must be measured, not assumed harmless.

What the Eden patch looks like

The selected method is a declarative method/path policy on the write routes of the affected virtual path. The following is a readable sketch of the rule, not the deployed schema:

{
  "version": 1,
  "profile": "http-structure-v1",
  "policy": {
    "version": 1,
    "rules": [
      { "scope": { "methods": ["PUT"], "paths": ["/uploads/*"], "decision": "reject" }
    ]
  }
}

Qualification must pin the exact deployment: which write paths exist, whether partial PUT is enabled, whether the session file store is active, and which business uploads must survive. Direct backend access and alternate write routes (JSP deployment paths, manager application) are separate coverage questions recorded in the tracking issue.

A simple API deployment with Eden

Prerequisite: an already-installed Eden foundation with the HTTP remediation adapter enabled, an operator with CONFIGURE authority on the target interlay, and a pinned deployment configuration showing the affected write/partial-PUT path. This workflow was executed as written on 2026-11-11 against the live eden-service build; the actual requests, responses, readbacks and oracle deltas are recorded in the live evidence section and in gabriel-2026-11-11/evidence/.

1. Select the patch. The immutable reviewed PUT /uploads deny policy shown above was selected as the interlay's settings.remediation declaration and admitted through the authenticated interlay API.
2. Deploy through the API. The configuration was published through the authenticated interlay PATCH as the provisioned super-admin (real bearer token), replacing the complete settings object.
3. Confirm it is effective. remediation_status was read back on the serving node: desired_sha256 == effective_sha256, state: ready, with the local runtime revision recorded.
4. Verify the outcome. The same partial-PUT exploit sequence was replayed through the interlay: the write was refused with no file placed and no session deserialization, while required uploads/downloads and application writes continued, observed independently at Tomcat.

Results and limits

The patch removes the file-placement prerequisite for the declared route — remediation, not repair: the patched Tomcat remains the durable fix, and the deserialization defect itself is untouched. A path that does not traverse Eden, or an alternate write route outside the policy scope, remains a coverage hole.

Live boundary evidence — 2026-11-11 UTC (run gabriel-live)

Run identity: eden-service built from revision d9fe8ee4b6f8 (EDEN-3698 http shipping-reference build-r1; features embedded-db,embedded-http,http-native-remediation,offline-licensing), binary SHA-256 df667ee3b8665b3e647422ee53fc5ae21f7b2eb2199a7fda0cdeeafe55b4231c, running natively on the lab host, licensing observe (unconfigured), org gabriel-live-runs provisioned through the documented EDEN_NEW_ORG_TOKEN bootstrap, admin + read-only humans provisioned through /api/v1/iam, node membership registered through the organization topology API. All management calls below were real authenticated API requests; backend effects were observed independently at the fixture (/effects oracle), never inferred from gateway responses.

Executed protocol (evidence gabriel-2026-11-11/evidence/tomcat.json.json):

Case-specific observations (Tomcat)

• Fixture: write-enabled uploads (PUT /uploads/{name}) where a *.session filename triggers the (simulated) deserialization sink. After activation the partial PUT was refused 403 with no deserialization effect; GET /index.html and an ordinary POST upload continued to work through the gateway.
• Policy: http-structure-v1 deny scoped to PUT under /uploads.
• Honest limits: the tested shape is the partial PUT on the declared route; alternate write routes and direct backend access remain coverage questions of the case. Blocks the file-placement prerequisite, not the deserialization defect itself.

Two-tier remediation — policy vs native module (live, installed 1.103.0, 2026-11-11)

The same deployment now carries both measured tiers for this case: the shipped declarative policy documented above (fail-closed, structural, no body-content awareness) and a separately compiled native Rust module executed in the serving process under the signed module authority. Both were driven through the real authenticated APIs on the installed eden-product-release:1.103.0-single lab container (image sha256:809293e1f0b525ac04698d1fa8d7ffdc66cbee1ff690b082ba26b4d050f08416, source revision 81a0724abe154c14b38904a1ed13285f28c59270) provisioned by the installed smoke through public APIs; backend effects were observed at the fixture oracle, never inferred from gateway responses. Evidence: gabriel-2026-11-11/module2/evidence/tomcat-module.json.

Tier 2 — what the module does

gabriel_tomcat_v1 inspects PUT|POST /uploads/* request lines and refuses only upload names colliding with the serialized-session store (.session suffix) — the deserialization sink's placement prerequisite. Every other upload name passes.

Module gabriel_tomcat_v1, binary SHA-256 497a1a6879ba2dd9f11d0ed7d8750bc5faba71a7bd1febbddaa9caad94c93892 (source 5123699e54ef8f70a2cab472875d92660ec6bad4b4ea511368695d217fbaa871).

Measured comparison

Tier 1 must disable uploads entirely; tier 2 refuses exactly the store-collision names and keeps the write capability. The write-path prerequisite removal (uploads of session-colliding names) is what both tiers enforce; an alternate route outside the interlay remains a stated coverage hole.

Performance trace (Eden telemetry, live 2026-11-11)

Measured with Eden's built-in request telemetry (gateway.request.baseline spans, read back through /api/v1/analytics/telemetry?signal=traces), not client wall-clock. Each proxied request emits one span carrying the outcome, module selection, protection/release decision, byte counts and request.duration_us. Arms: normal request through an unprotected interlay; the identical benign request through the tier-2 module interlay (77151bf0…, publication 9c3934fa…); and the attack under the module. Keep-alive connections, warm requests only.

• normal: ('bypass', 'not_selected', 'false', 'false')×23
• remediated benign: ('allowed', 'module', 'true', 'true')×23
• attack: ('refused', 'module', 'true', 'false')×14

Refusal-path latency (attack under remediation): mean 137 µs (p50 118 µs, p95 231 µs).

Eden end-to-end remediation latency (in-product): benign requests mean 384 µs (p50 359 µs, p95 528 µs). Attack/refusal arm: mean 137 µs, p95 231 µs.

Sample benign span under the module (attributes as recorded by Eden):

  `request.outcome`: `allowed`
  `request.selection`: `module`
  `request.protected`: `true`
  `request.released`: `true`
  `request.input_bytes`: `150`
  `request.forwarded_bytes`: `150`
  `request.response_bytes`: `197`
  `request.duration_us`: `329`
  `gateway.publication`: `<lab publication ID>`
  `gateway.generation`: `4`

Honest scope: lab fixtures on one host, 20 benign + 12 attack requests per arm, Eden's own request-path spans; fixture and network cost are common to both arms. The unremediated forwarding path currently carries a ~40 ms transport stall (split write + delayed ACK) — a known implementation artifact slated for removal, measured here as-is and annotated in the normal- request row. Not the EDEN-3640 performance qualification contract.

(Trace section generated from module2/telemetry-ab/tomcat-telemetry.json; module reactivated shards 32, operation f0f38a45….)

Performance trace (Eden telemetry, live 2026-11-11)

Measured with Eden's built-in request telemetry (gateway.request.baseline spans, read back through /api/v1/analytics/telemetry?signal=traces), not client wall-clock. Latency Eden introduces per remediation arm for this case:

Eden-added latency for the tier-2 remediation: benign requests complete in 384 µs mean (p50 359 µs, p95 528 µs) in-product, module evaluation included. Honest scope: lab fixtures on one host, fresh connection per measured request, warm requests only; module reactivated shards 2/2 with a fresh signed statement per measurement; spans carry the full authority lineage (module SHA-256, operation, publication). Not the EDEN-3640 performance qualification contract.

Tested against: eden-product-release:1.103.0-single (image sha256:809293e1f0b525ac…, build target eden_service_single, tooling revision 81a0724a…). Raw data: module2/telemetry-ab/tomcat-telemetry.json and module2/tier1-ab/tomcat-tier1.json.

Authority lifecycle and negatives

The module reached the serving process only through the full signed authority lifecycle (module2/evidence/tomcat-module.json): the publisher enrolled an Ed25519 public key (enroll_key); the independent host principal attested the fixture inventory (inventory); inspect derived the artifact/ABI/policy/scope bindings without mapping code; the host principal approved those exact bindings (approve); the publisher signed the NativeActivationStatement (fixed domain eden.module-admission.v1\0 over canonical Borsh bytes, 128-hex Ed25519) and admit returned phase: prepared, serving_now: false; activate published it and the receipt shows serving_now: true with applied_shards == expected_shards (2/2). Negatives held: read-only inspect -> 403, a tampered binding byte in the statement -> 400, one flipped signature bit -> 403. Revoking the operation made the interlay fail closed (503 E1A26; nothing forwarded, zero backend effects) rather than silently unprotected, and a fresh host approval plus newly signed activation restored surgical protection (the session-collision PUT was refused 403 again).

Honest scope

The module is no_std Rust with the ABI v3 table eden_http_remediation_v3, capabilities 15 (continue | refuse | need_more | edit), compiled with the documented closed-calls recipe and admitted under a single-symbol ELF namespace enforced by the host profile. Its rule keys only on the the upload-name segment of the request line and its .session suffix; it is a lab-grade demonstration of content-aware remediation on this delivery shape, not a production Tomcat patch, and the vendor's patched release remains the durable repair.

Cost and scalability record (2026-11-11 session)

Session ledger (heuristic: chars/4 tokens; GLM-5.3 pricing $0.0699/M input, $1.40/M output): the portfolio session closed at **$14.05** (200.3M input / 34.4k output tokens across 895 assistant turns). The tier-2 native-module batch of seven case studies cost **$9.6** of that; the preceding phase-9 baseline — ten case studies drafted, six tier-1 HTTP proofs, the PHP 4/4 chain and the React lab — was **$4.46**. Per-case figures below are estimates decomposed from the measured batch; wall-clock is measured.

Case-specific content: request-line name/suffix matching. Case-specific debugging: one round (suffix must exclude the HTTP/1.1 version token).

Scalability. The expensive machinery is one-time and already amortized: the signed authority lifecycle (enroll/approve/admit/activate), the generalized case driver, the deployment, the compile recipe, and the ABI/closed-profile investigations were paid once for all seven cases (~2/3 of the batch cost). Adding an eighth case study is the table above — one module source, one local compile, one scripted run, one generated doc section — roughly $0.5-0.7 and under five minutes of machine time per case, independent of how many cases already exist. That flat marginal cost is the scalability claim: portfolio growth is linear in cases, not compounding, and the per-case rule is bounded (97 lines of Rust keyed to one delivery shape).

Evidence appendix

• Tracking issue EDEN-3167
• API freeze and portfolio contract EDEN-3160
• Skill references: api, testing

Decision flow

The source checks the selected upload routes and suffix. This illustrates the matcher’s decision, not every Tomcat session-file placement path.

Complete normalized policy declaration

This is the full settings.remediation example included in the reader-facing case study. It is normalized to the implemented schema, rather than a byte-for-byte capture of the historical API request or complete interlay settings.

{
  "version": 1,
  "profile": "http-structure-v1",
  "policy": {
    "version": 1,
    "rules": [
      {
        "scope": {
          "methods": [
            "PUT"
          ],
          "path": "/uploads",
          "descendants": true
        },
        "require": {
          "kind": "deny"
        }
      }
    ]
  }
}

Complete native module source

The complete available source for gabriel_tomcat_v1 is reproduced below from the reviewed repository snapshot. The source SHA-256 is 5123699e54ef8f70a2cab472875d92660ec6bad4b4ea511368695d217fbaa871. This identifies the source bytes, rather than independently proving the binary build or the recorded test results. The case’s experiment status and limitations still apply, including any failed import, partial coverage, or withdrawn claim. Original source comments are preserved verbatim.

//! Gabriel case module: session-store collision refusal for CVE-2025-24813.
//!
//! Tier-2 remediation for the Tomcat CVE-2025-24813 delivery shape: on the
//! upload routes, refuse only writes whose upload name collides with the
//! serialized-session store (`.session` suffix — the deserialization sink).
//! Ordinary uploads of any other name continue unchanged, preserving the
//! write capability (zero false positives). This no_std module owns no
//! storage, allocator, background work, or persistent state.

#![no_std]

use core::mem::size_of;
use core::panic::PanicInfo;
use eden_gateway_remediation::message::{Availability, MessageInput};
use eden_gateway_remediation::module::{
    ApiHeader, CONTINUE, Decision, EDIT, EditSink, Edits, Failure, ModuleOutput, NEED_MORE, REFUSE, RemediationModule,
};
use eden_gateway_remediation::native_http_v3::{self as http, Api, Input, Request};

/// Upload routes in scope (the write-enabled default servlet surface).
const ROUTES: [&[u8]; 2] = [b"PUT /uploads/", b"POST /uploads/"];

/// The store-collision suffix this module refuses on those routes.
const SESSION_SUFFIX: &[u8] = b".session";

/// Stateless rule: refuse .session-named writes on upload routes.
struct RefuseSessionCollision;

fn request_line(head: &[u8]) -> &[u8] {
    let end = head.iter().position(|&b| b == b'\n').unwrap_or(head.len());
    let end = if end > 0 && head[end - 1] == b'\r' { end - 1 } else { end };
    &head[..end]
}

impl RemediationModule<Request<'_>> for RefuseSessionCollision {
    /// Wait for the complete head, then refuse only session-collision names.
    fn evaluate(&self, request: &Request<'_>, edits: &mut Edits<'_>) -> Result<Decision, Failure> {
        let _ = edits;
        let _ = match request.header_count() {
            Availability::Ready(count) => count,
            Availability::NeedMore => return Ok(Decision::NeedMore),
            Availability::Unavailable => return Err(Failure),
        };
        let message = request.message();
        if !message.is_complete() {
            return Ok(Decision::NeedMore);
        }
        let head = message.prefix();
        let line = request_line(head);
        let mut matched = false;
        for route in ROUTES {
            if line.starts_with(route) {
                matched = true;
                break;
            }
        }
        if !matched {
            return Ok(Decision::Continue);
        }
        // The upload name is the path segment after /uploads/, up to the space
        // that starts the HTTP-version token (never part of the name).
        let rest = &line[b"/uploads/".len()..];
        let name_len = rest.iter().position(|&b| b == b' ').unwrap_or(rest.len());
        if rest[..name_len].ends_with(SESSION_SUFFIX) {
            return Ok(Decision::Refuse);
        }
        Ok(Decision::Continue)
    }
}

static API: Api = Api {
    header: ApiHeader { version: http::VERSION, table_size: size_of::<Api>() as u32 },
    input_size: size_of::<Input>() as u32,
    output_size: size_of::<ModuleOutput>() as u32,
    sink_size: size_of::<EditSink>() as u32,
    message_size: size_of::<MessageInput>() as u32,
    state_bytes: 0,
    capabilities: CONTINUE | REFUSE | NEED_MORE | EDIT,
    required_inputs: 0,
    evaluate: Some(evaluate),
};

unsafe extern "C" fn evaluate(input: *const Input, sink: *const EditSink, output: *mut ModuleOutput) -> i32 {
    unsafe { http::invoke(&RefuseSessionCollision, input, sink, output) }
}

#[unsafe(no_mangle)]
pub extern "C" fn eden_http_remediation_v3() -> *const Api {
    &API
}

#[panic_handler]
fn panic(_information: &PanicInfo<'_>) -> ! {
    loop {
        core::hint::spin_loop();
    }
}