CVE-2025-66411
NixOS vulnerability analysis and mitigation

Overview

CVE-2025-66411 is an information disclosure vulnerability in Coder (the remote development environment platform) where Workspace Agent manifests containing sensitive environment variables were logged in plaintext without sanitization. It affects all Coder versions prior to 2.26.5, 2.27.0–2.27.6 (fixed in 2.27.7), and 2.28.0–2.28.3 (fixed in 2.28.4). The vulnerability was disclosed on December 3, 2025, via a GitHub Security Advisory. The CVSS v3.1 base score is 5.5 (Medium) per NVD, though the GitHub advisory rates it 7.8 (High) (GitHub Advisory).

Technical details

The root cause is classified as CWE-532 (Insertion of Sensitive Information into Log File). In agent/agent.go, the handleManifest function logged the full Workspace Agent manifest proto object — including sensitive environment variables — via a.logger.Info(ctx, "fetched manifest", slog.F("manifest", mp)) without any sanitization or redaction. By default, Workspace Agent logs are directed to stderr (configurable via CODER_AGENT_LOGGING_HUMAN), meaning these logs could be captured by any process or user with local access to the workspace VM or Kubernetes pod, or by downstream log aggregation systems such as SIEMs. The fix, implemented in PR #20968, simply removes the manifest field from the log call, changing the line to a.logger.Info(ctx, "fetched manifest") (GitHub Advisory, Fix Commit).

Impact

Successful exploitation allows a local attacker with limited access to a Coder Workspace (VM, Kubernetes Pod, etc.) or access to a connected logging system (SIEM, log aggregation stack) to read sensitive environment variables from the agent manifest logs in plaintext. The confidentiality impact is high, as the exposed data may include API keys, secrets, tokens, or other credentials injected into the workspace environment. Depending on the sensitivity of the exposed values, this could enable lateral movement, privilege escalation, or unauthorized access to other systems or services (GitHub Advisory).

Exploitability

No public exploit code or active in-the-wild exploitation has been reported for this vulnerability. The EPSS score is approximately 0.012% (0.000120), indicating a very low probability of exploitation in the near term. The vulnerability is not listed in the CISA Known Exploited Vulnerabilities (KEV) catalog. Exploitation requires low-privilege local access to the affected workspace environment or access to the logging infrastructure, limiting the attacker pool (GitHub Advisory, Feedly).

Exploitation steps

  1. Gain local access: Obtain low-privilege access to a Coder Workspace (e.g., a VM or Kubernetes Pod) running a vulnerable version of the Coder agent (prior to 2.26.5, 2.27.7, or 2.28.4), or gain read access to the workspace's log output or a connected SIEM/logging stack.
  2. Locate agent logs: By default, Workspace Agent logs are written to stderr. Identify where these logs are captured — this may be a container log (e.g., kubectl logs <pod>), a systemd journal, a file, or a third-party log aggregator.
  3. Search for manifest log entries: Look for log lines containing "fetched manifest" with a manifest field, which will contain the full agent manifest proto object in plaintext.
  4. Extract sensitive values: Parse the manifest log entry to extract environment variables and other sensitive configuration values (e.g., API keys, tokens, secrets) that were injected into the workspace by Terraform or platform operators.
  5. Leverage extracted credentials: Use the obtained secrets to access other systems, escalate privileges, or move laterally within the organization's infrastructure (GitHub Advisory, Fix PR).

Indicators of compromise

  • Logs: Presence of log entries containing "fetched manifest" with a manifest field (structured log key-value) in Workspace Agent stderr output or aggregated logs — this indicates a vulnerable version is running and sensitive data may have been exposed.
  • Logs: Unexpected queries or searches in SIEM/log aggregation systems targeting Coder agent log streams, particularly filtering for manifest or environment variable content.
  • Process: Workspace agent processes (on VM or Kubernetes Pod) running Coder versions prior to 2.26.5, 2.27.7, or 2.28.4, identifiable via version flags or binary metadata.
  • Network: Unusual outbound connections from workspace environments to external IPs shortly after workspace provisioning, which may indicate credential abuse following log exfiltration (GitHub Advisory).

Mitigation and workarounds

Coder has released patched versions: 2.26.5, 2.27.7, and 2.28.4. Platform operators should upgrade to one of these versions as the primary remediation. As a temporary workaround, Workspace Agent logging can be disabled by setting the environment variable CODER_AGENT_LOGGING_HUMAN=/dev/null, which prevents sensitive data from being written to logs; however, this also suppresses all human-readable agent logs. Operators should also review and rotate any secrets or credentials that may have been exposed in agent logs prior to patching (GitHub Advisory).

Additional resources


SourceThis report was generated using AI

Related NixOS vulnerabilities:

CVE ID

Severity

Score

Technologies

Component name

CISA KEV exploit

Has fix

Published date

CVE-2026-85706CRITICAL10
  • GitLab logoGitLab
  • gitlab
YesYesSep 12, 2026
CVE-2026-88009HIGH8.8
  • NixOS logoNixOS
  • github.com/traefik/traefik/v2
NoYesSep 10, 2026
CVE-2026-88008HIGH7
  • NixOS logoNixOS
  • github.com/traefik/traefik/v2
NoYesSep 10, 2026
CVE-2026-88012MEDIUM5.3
  • NixOS logoNixOS
  • traefik-fips-3
NoYesSep 10, 2026
CVE-2026-88011MEDIUM5.3
  • NixOS logoNixOS
  • github.com/traefik/traefik/v2
NoYesSep 10, 2026

Free Vulnerability Assessment

Benchmark your Cloud Security Posture

Evaluate your cloud security practices across 9 security domains to benchmark your risk level and identify gaps in your defenses.

Request assessment

Get a personalized demo

Ready to see Wiz in action?

"Best User Experience I have ever seen, provides full visibility to cloud workloads."
David EstlickCISO
"Wiz provides a single pane of glass to see what is going on in our cloud environments."
Adam FletcherChief Security Officer
"We know that if Wiz identifies something as critical, it actually is."
Greg PoniatowskiHead of Threat and Vulnerability Management