Back to skill

Security audit

Finix JS

Security checks for vulnerabilities and agentic risk

Overview

This is a payment-integration guide with disclosed Finix sandbox examples; it has a credential-handling documentation risk but no hidden code or off-purpose behavior.

Reasonable to install for Finix payment integration work, but treat the curl credentials as sandbox-only examples. Do not place live Finix credentials in browser code, shell history, logs, CI output, or copied command lines; keep production credentials server-side in a secret manager or protected configuration.

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

Warning
Location
SKILL.md:518
Finding

Hard-Coded Finix Sandbox API Credentials in Executable Examples

Content
View full analysis

Vulnerability Details

File Location: SKILL.md, lines 518–520, 551–553, and 590–592
Vulnerability Type: Hard-coded API credentials and unsafe command-line secret handling
Risk Level: Medium

Vulnerable Code

The same HTTP Basic Auth credential is embedded in three executable curl examples:

bash
curl -i -X POST \
  -u USfdccsr1Z5iVbXDyYt7hjZZ:313636f3-fac2-45a7-bff7-a334b93e7bda \
  https://finix.sandbox-payments-api.com/identities \
bash
curl -i -X POST \
  -u USfdccsr1Z5iVbXDyYt7hjZZ:313636f3-fac2-45a7-bff7-a334b93e7bda \
  https://finix.sandbox-payments-api.com/payment_instruments \
bash
curl -i -X POST \
  -u USfdccsr1Z5iVbXDyYt7hjZZ:313636f3-fac2-45a7-bff7-a334b93e7bda \
  https://finix.sandbox-payments-api.com/transfers \

The document later states that these are public Finix sandbox test credentials. That limits their apparent production impact, but does not eliminate the unsafe pattern.

Technical Analysis

The Skill embeds a realistic username/password pair directly in copyable commands despite declaring FINIX_USERNAME and FINIX_PASSWORD as environment requirements. Anyone who can read the Skill can retrieve and use the pair. If the credentials remain active, they may permit authenticated operations against the Finix sandbox API, including creating identities, claiming tokens as payment instruments, and submitting test transfers within the permissions and resources associated with that sandbox account.

The curl -u username:password pattern also encourages users to place credentials in shell command arguments. Commands are commonly retained in shell history, terminal transcripts, CI logs, support records, and audit telemetry. Depending on the operating system and timing, command arguments may also be visible to other local users through process inspection. Users who replace the example pair with production credentials could therefore expose materially sensitive secrets even though the embedded pair ...[truncated 2027 chars]

Remediation
View remediation

Remediation Suggestions

  1. Remove the literal username/password pair from every example and revoke or rotate it if it was not intentionally designed for unrestricted public use.
  2. Use clearly nonfunctional placeholders in documentation, such as:
    bash
    curl --user "${FINIX_USERNAME}" \
      https://finix.sandbox-payments-api.com/identities
    
  3. Warn that environment-variable expansion still places the expanded secret in the process argument list for many command-line patterns. Prefer a protected configuration file, secret manager, or server-side SDK mechanism that does not expose credentials in command arguments.
  4. Ensure credential-bearing files have restrictive permissions, are excluded from version control, and are never printed in application, CI, or HTTP logs.
  5. Use separate sandbox and production credentials, grant only the API permissions required by the integration, and rotate credentials periodically.
  6. Add explicit documentation that browser code must never receive FINIX_USERNAME or FINIX_PASSWORD; only the public Application ID and Merchant ID should be browser-visible.
  7. Replace realistic tokens, identity IDs, payment-instrument IDs, and merchant IDs with unmistakably synthetic placeholders where executable realism is unnecessary.
  8. Add automated secret scanning to repository and publication workflows to reject credential-like values before release.
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)

External Transmission

Medium
Category
Data Exfiltration
Confidence
60% confidence
Finding

Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

Content

Scanner excerpt · SKILL.md (reported line 518)May include surrounding context.

You can create a fully-populated Identity, or an empty one with just identity_roles and type — both are valid. Populating it up front avoids asking the buyer for the same info later.

bash
curl -i -X POST \
  -u USfdccsr1Z5iVbXDyYt7hjZZ:313636f3-fac2-45a7-bff7-a334b93e7bda \
  https://finix.sandbox-payments-api.com/identities \
  -H 'Content-Type: application/json' \

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
88% confidence
Finding

The example config sets language: 'en' for Plaid Link, which forces an English locale in a user-facing payment flow. The document does not mention that this is optional, user-selectable, or limited to an English-speaking deployment context.

Content

No source excerpt is available for this finding.