Back to skill

Security audit

Tokenburg

Security checks for vulnerabilities and agentic risk

Overview

This skill is coherent for playing Tokenburg, but it routes a sensitive bearer token to a configurable server without enough safeguards or warning.

Install only if you trust the configured Tokenburg server. Treat TOKENBURG_BEARER like a password, do not share it, and avoid setting TOKENBURG_SERVER away from https://tokenburg.com unless you are intentionally using a trusted test server with a separate token.

Vulnerability Patterns
  • Insecure Skill Coding PracticesFinds exploitable flaws such as hardcoded secrets or command injection
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
Findings (1)

T09 · Insecure Skill Coding Practices

Error
Location
SKILL.md:55
Finding

Bearer Token Exfiltration Through a Configurable Server Origin

Content
View full analysis

Vulnerability Details

File Location: SKILL.md, lines 55–68
Vulnerability Type: Credential disclosure through an unrestricted network destination
Risk Level: High

Vulnerable Code

text
- `TOKENBURG_SERVER` (optional, default `https://tokenburg.com`) — the server.

Your **character ID is the middle segment of the Bearer token** (between the leading
`agt_` and the trailing `_<random>`). Below, read `<server>` as `TOKENBURG_SERVER` and
`<character ID>` as that character ID.

---

## 1. Authentication

Send this header on every HTTP request:

Authorization: Bearer

text

Technical Analysis

The Skill requires the Agent to send the owner-provided TOKENBURG_BEARER credential with every HTTP request. At the same time, it permits the request origin to be selected through the optional TOKENBURG_SERVER environment variable.

The instructions do not require the configured server to:

  • Match the official https://tokenburg.com origin.
  • Use HTTPS.
  • Belong to an explicit allowlist.
  • Receive informed user approval before credentials are sent.
  • Preserve the trusted origin across redirects.

Sending the bearer token to the official Tokenburg service is necessary for the declared functionality. Sending it automatically to an arbitrary configurable origin is not necessary and exceeds the minimum privilege required.

An actor capable of setting or influencing TOKENBURG_SERVER can direct authenticated requests to a server they control. Because the Skill explicitly instructs the Agent to attach the bearer token to every request, the malicious endpoint can capture the credential. This finding concerns the unsafe credential-routing design; the audited file does not itself establish who can modify the environment.

Attack Path

  1. An attacker gains the ability to set or influence TOKENBURG_SERVER in the Skill's execution environment.
  2. The attacker configures it to an endpoint under their control, such as `https://atta ...[truncated 1416 chars]
Remediation
View remediation

Remediation Suggestions

  1. Pin credentialed requests to the official origin

    • Permit TOKENBURG_BEARER to be sent only to the exact normalized origin https://tokenburg.com.
    • Validate the URL scheme, hostname, and effective port before constructing any authenticated request.
  2. Require HTTPS

    • Reject http:// origins for all requests carrying the bearer token.
    • Do not permit TLS validation to be disabled.
  3. Do not forward credentials across redirects

    • Disable automatic redirects for authenticated calls or revalidate the destination after every redirect.
    • Strip the Authorization header whenever the origin changes.
  4. Separate development credentials

    • If custom servers are necessary for development, require explicit informed user approval and use separate, narrowly scoped test credentials.
    • Never reuse production bearer tokens with development or user-supplied origins.
  5. Use an explicit allowlist

    • Replace unrestricted TOKENBURG_SERVER support with a documented allowlist of trusted origins.
    • Compare normalized origins rather than relying on hostname substring or suffix checks.
  6. Reduce credential scope and lifetime

    • Use short-lived, revocable tokens with only the API permissions necessary for the character.
    • Provide token rotation and immediate revocation after suspected disclosure.
  7. Clarify the Skill instructions

    • State explicitly that the bearer token must never be transmitted to arbitrary URLs, image hosts, redirects, or origins obtained from remote API content.
    • Require confirmation before any exceptional credential destination is used.
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Findings (2)

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
86% confidence
Finding

This markdown file instructs the skill to use the TOKENBURG_BEARER environment variable and explains how to derive the character ID from it, but it does not include any warning that the bearer token is a sensitive credential that must not be exposed or shared. Because the skill makes authenticated network requests on the user's behalf, the description should clearly disclose the sensitivity of this credential and caution against leaking it in outputs or logs.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

The documentation requires amend_constitution to use Japanese article labels such as 第N条, 第三条, and 附則. This imposes a specific language/locale convention in the skill instructions without offering the user a choice or clearly justifying the constraint as a region-specific compliance requirement.

Content

No source excerpt is available for this finding.