Skip to main content

Loom for AWS EUVDEUVD-2026-91773

| CVE-2026-103958 HIGH
Server-Side Request Forgery (SSRF) (CWE-918)
2026-10-02 AMZN
8.3
CVSS 4.0 · Vendor: AMZN
View

Attack ChainAIDerived

Hypothetical attack flow derived from CVE metadata – not a confirmed exploit chain

  1. Access
    Authenticate as privileged Loom user
  2. Delivery
    Register/update/test tool server with crafted internal address
  3. Exploit
    Backend performs server-side fetch of metadata endpoint
  4. Execution
    Response body echoed to attacker
  5. Impact
    Harvest container IAM role credentials and internal data

Vulnerability Assessment

Weakness
CWE
OWASP Top 10:2025
A01:2025 - Broken Access Control
ASVS 5.0
V4 - API and Web Service URL allowlisting, SSRF prevention
MITRE ATT&CK

2 techniques across 2 tactics: Initial Access, Credential Access

The technique list with ATT&CK links and the kill-chain view is part of the Basic plan and up.

Exploitation Conditions

Exploitation requires an authenticated account with privileges to register, update, or test a tool server or remote agent (the MCP Server / A2A Agent integration feature) - consistent with PR:H. The attacker supplies a crafted connection address (e.g., the AWS instance/container metadata endpoint at 169.254.169.254 or an internal host); the server-side fetcher then follows redirects and echoes the response body back, which is the mechanism enabling credential and internal-data disclosure. Limiting factors: no anonymous/pre-auth path (privileged integration-management access is mandatory), exploitation depends on the container running with an IAM role worth stealing, and deployments on or upgraded to 1.7.0 (which adds the IP-validated, DNS-pinned fetcher and connection-sink guards) are not affected.

Impact

This is a credible, high-priority finding for anyone actually running Loom for AWS, but the real-world risk is bounded by its prerequisites. The CVSS 4.0 vector (AV:N/AC:L/AT:N/PR:H/UI:N) describes a network-reachable, low-complexity attack that nonetheless requires HIGH privileges (PR:H) - the actor must already be an authenticated user with rights to register, update or test a tool server or remote agent, so this is not a pre-auth internet mass-exploitation bug. Its severity comes from impact, not accessibility: VC:H plus SC:H reflect that the SSRF can reach the AWS instance/container metadata endpoint and disclose the application's own container IAM role credentials, enabling lateral movement into the surrounding AWS account - a classic SSRF-to-cloud-credential-theft escalation (CWE-918). No EPSS score was provided and the CVE is not in CISA KEV; there is no public exploit identified at time of analysis, so exploitation probability signals are absent rather than reassuring. There is no conflict between the CVSS vector and the description - both agree on network vector plus authenticated/privileged access. The practical priority is HIGH for Loom operators (credential theft is confirmed by the vendor advisory and the fix) but is naturally scoped by the small deployment footprint and the privileged-insider or compromised-account precondition.

Exploit ScenarioDerived

An authenticated user (or an attacker who has phished/compromised a privileged Loom account) with permission to manage integrations registers or 'tests' a tool server / remote agent, supplying a crafted connection address pointing at the AWS metadata service or an internal host. The backend fetches that address server-side, and because responses are echoed back (and redirects followed), the attacker reads the container role's temporary IAM credentials and probes otherwise-unreachable internal endpoints. No public exploit code is identified at time of analysis, but the technique is well-understood and the trigger is a normal product feature.

This Week Schedule remediation this week. Why: HIGH severity (CVSS 7.0-8.9).
Patch available 0.3% exploitation probability (EPSS)

Get an e-mail when this CVE is exploited, patched or re-scored. Free account.

Alert me

AnalysisAIv1 – Oct 02, 2026

Server-side request forgery in Loom for AWS before 1.7.0 lets an authenticated user with privileges to register, update, or test a tool server or remote agent (the MCP Server / A2A Agent integration feature) supply a crafted connection address that the backend fetcher follows and whose response body it echoes back, exposing the application container's own IAM role credentials and responses from arbitrary internal network locations. Exploitation requires a high-privileged authenticated account (CVSS PR:H) and is only meaningful where the container runs with a role worth stealing; there is no anonymous or pre-authentication path, and no public exploit identified at time of analysis. Deployments on or upgraded to 1.7.0, which adds IP-validated, DNS-pinned fetchers and connection-sink guards, are not affected.

RemediationAI

Upgrade Loom for AWS to version 1.7.0 or later - the vendor-released fix, available per the AWS security bulletin (https://aws.amazon.com/security/security-bulletins/2026-124-aws/) and the v1.7.0 release (https://github.com/awslabs/loom/releases/tag/v1.7.0); 1.7.0 adds IP-validated, DNS-pinned fetchers and guards the MCP/A2A connection sinks against redirect-based SSRF and response-body echo, so there is no partial-patch state to reason about - move to the fixed build. If immediate upgrade is not possible, apply compensating controls knowing their trade-offs: restrict the IAM permissions that allow users to register, update, or test tool servers and remote agents to the smallest trusted set (this reduces the attack surface but may block legitimate integration work), and audit/reduce the privileges of the container's own IAM role so that stolen role credentials yield little (this is least-privilege hygiene rather than a fix and may break workloads that genuinely need those permissions). Enforce IMDSv2 with hop limits on the instance/container metadata endpoint and block egress from the Loom backend to link-local (169.254.0.0/16) and sensitive internal ranges at the network layer; this defeats the credential-theft path but can break legitimate internal integrations Loom relies on, so it should be scoped and tested. Rotate any container-role credentials and internal secrets that may have been exposed before upgrading.

Affected VersionsENISA

loom 0 <1.7.0

Affected Packages

euvd loom fix: 1.7.0

Source: OSV (Open Source Vulnerabilities)

Affects your stack

Affects 1 package in euvd

Checking this CVE against your own stacks and servers is part of the Basic plan and up.

Technical ContextAI · show details

The vulnerability is a classic server-side request forgery (CWE-918) rooted in Loom for AWS's integration plumbing for Model Context Protocol (MCP) servers and A2A agents. When a user registers, updates, or tests a tool server or remote agent, the backend performs an outbound HTTP fetch to the user-supplied connection address; the pre-1.7.0 fetcher neither validated the destination IP nor pinned DNS resolution, and it followed redirects and returned the fetched response body to the caller. That combination lets the request be steered at link-local metadata services (e.g., 169.254.169.254) or arbitrary internal hosts, with the response reflected back through the application - turning a legitimate integration-management workflow into an internal-network read primitive and a credential-disclosure primitive for the container's own IAM role. The 1.7.0 release notes confirm the fix scope: connection sinks for MCP/A2A are now guarded against SSRF via redirect and response-body echo, and the same IP-validated, DNS-pinned fetcher was extended to every other outbound OAuth2/OIDC call in the backend (OIDC discovery, JWKS, generic OAuth2, per-user authorizer linking, Cognito, and the login-callback token exchange), closing a residual SSRF/token-disclosure gap where a deployment-trusted issuer host is now required before any token exchange sends a client secret or user access token.

Know when Loom gets its next vulnerability

  • Watch products and get an email when something new is published
  • Check this CVE against your own stack and servers
  • Exploit and PoC references, threat intelligence from 22 sources

No credit card · 7-day full trial

Severity by source

Vendor (AMZN) PRIMARY 8.3 HIGH
CVSS:4.0/AV:N/AC:L/AT:N/PR:H/UI:N/VC:H/VI:N/VA:N/SC:H/SI:L/SA:N
vuln.today AI 7.6 HIGH

Network-reachable integration feature (AV:N/AC:L) but requires privileged integration-management rights (PR:H); SSRF reaches the metadata/internal systems so Scope changes (S:C) with high confidentiality and only minor integrity impact.

3.1 AV:N/AC:L/PR:H/UI:N/S:C/C:H/I:L/A:N
4.0 AV:N/AC:L/AT:N/PR:H/UI:N/VC:H/VI:N/VA:N/SC:H/SI:L/SA:N

Primary rating from Vendor (AMZN).

CVSS VectorVendor: AMZN

Attack Vector
Network
Attack Complexity
Low
Attack Requirements
None
Privileges Required
High
User Interaction
None
Vulnerable system: Confidentiality
High
Vulnerable system: Integrity
None
Vulnerable system: Availability
None
Subsequent system: Confidentiality
High
Subsequent system: Integrity
Low
Subsequent system: Availability
None

Lifecycle Timeline

2
CVSS changed
Oct 02, 2026 - 19:22 NVD
7.6 (HIGH) 8.3 (HIGH)
CVE Published
Oct 02, 2026 - 19:07 cve.org
HIGH 8.3

DescriptionCVE.org

Server-side request forgery in the tool server and remote agent connection handling in Loom for AWS before 1.7.0 might allow an authenticated remote user to obtain the credentials of the application's own container role and to read responses from arbitrary internal network locations, via a crafted connection address supplied when registering, updating or testing a tool server or remote agent.

To remediate this issue, users should upgrade to version 1.7.0 or later.

Technical POC Analysis

CWE-918 (Server-Side Request Forgery).

Full technical POC analysis - root cause, step-by-step exploitation and patch breakdown - is part of the Pro and Business plans.

Executive ViewAI

Why It Matters

A server-side request forgery flaw in Loom for AWS (versions before 1.7.0) allows a highly privileged authenticated user to coerce the backend into fetching attacker-chosen addresses and returning the responses, which can expose the application container's own cloud IAM role credentials and data from internal network services. Organizations running the affected tool-server or remote-agent integrations on AWS-hosted containers with attached roles face potential credential theft and lateral movement into adjacent internal systems. Because exploitation requires an already high-privileged account and a container role worth stealing, leadership should treat this as a privilege-escalation and cloud-credential exposure risk concentrated in a small number of administrative users rather than a broad external attack surface.

Business Impact: High
Compensating Controls

Apply vendor patch.

Threat Intelligence Overview

EPSS Score
0.3%
Top 76% exploitable
Prevalence
Low
new awslabs open-source AI-agent platform, limited production deployment footprint
Social Media
1
Sightings
Data Quality
6 sources
Confidence: High

How this decision was reached

The engine reads each signal below and applies a fixed list of rules from top to bottom. The first rule that matches sets the recommended action – like a statement, every line is shown.

SignalValue
Severity HIGH
CISA KEV No
Public POC No
EPSS 0.3%
Patch available Yes
Active threat No
Deciding rule
HIGH severity (CVSS 7.0-8.9)
What every action level means
Emergency Immediate action required. Criteria: CRITICAL + in CISA KEV (actively exploited) without patch, OR CRITICAL + EPSS > 30% without patch, OR CRITICAL + active threat signals without patch.
Act Now Prioritize within 24-48h. Criteria: in KEV with patch available, OR CRITICAL + public POC, OR HIGH + EPSS > 10%, OR CRITICAL + active threat, OR any CRITICAL severity.
This Week Schedule remediation this week. Criteria: HIGH severity (CVSS 7.0-8.9). Patch within normal change window.
This Month Plan within 30 days. Criteria: MEDIUM severity (CVSS 4.0-6.9), lower exploitation probability. Include in regular patch cycle.
Monitor Track and reassess. Criteria: LOW severity (CVSS 0.1-3.9), minimal exploitation risk. No immediate action needed.

SSVC AssessmentENISA

SSVC Decision
Scheduled
Exploitation
None
System Exposure· derived
Open
Automatable
No
Human Impact
High
Technical Impact
Partial

Priority Score

42
Low Medium High Critical
KEV: 0
EPSS: 0
CVSS: +42
PoC: 0
Weaponized: 0
Threat intel: 0
Gov advisory: 0

Attack Technique

Server-Side Request Forgery exploits applications that fetch remote resources based on user-supplied URLs.

This site uses cookies essential for authentication and security. No tracking or analytics cookies are used. Privacy Policy