Back to skill

Security audit

Bluepages Register

Security checks for vulnerabilities and agentic risk

Overview

The skill is a simple API guide, but it documents profile-changing and proxy-call features with weak or unclear controls that users should review before installing.

Install only if you are comfortable sending public agent profile details to DeepBlueBase and relying on its remote service controls. Before using update or proxy-call features, verify that profile changes require wallet signing or another real proof of ownership, and avoid proxying sensitive request data or untrusted/internal URLs unless the service documents strong restrictions.

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:57
Finding
Public Wallet Address Used as Profile-Update Authentication## Vulnerability Details **File Location**: `SKILL.md`, lines 57-68 **Vulnerability Type**: Authentication weakness caused by treating public wallet identifiers as credentials **Risk Level**: High ### Vulnerable Code Snippet ```markdown ## Update Your Profile ```bash PATCH https://api.deepbluebase.xyz/agents/{agent_id} Content-Type: application/json { "wallet_address": "0xYourBaseWallet", "description": "Updated description", "x_handle": "new_x_handle", "x402_endpoints": [ ... ] } ``` Only include fields you want to change. `wallet_address` is required for auth. ``` ### Technical Analysis The documented profile-update operation uses only `wallet_address` for authentication. A blockchain wallet address is a public identifier and does not prove that the requester controls the corresponding private key. The documented request contains no wallet signature, server-issued nonce, timestamp, expiring token, or equivalent proof-of-possession mechanism. If the remote API behaves exactly as documented, an attacker who knows an agent's public wallet address and agent ID can construct an update request that is indistinguishable from one sent by the wallet owner. The vulnerable behavior concerns the external service described by the skill; the reviewed project contains documentation only and no server-side implementation with which to verify whether undocumented controls exist. ### Attack Path 1. The attacker obtains a target's agent ID and public wallet address from its public directory entry or blockchain-related records. 2. The attacker sends a forged `PATCH` request to `/agents/{agent_id}`. 3. The attacker supplies the target's public wallet address in the `wallet_address` field. 4. The attacker replaces mutable fields such as `description`, `x_handle`, or `x402_endpoints`. 5. Users or autonomous agents relying on the directory may subsequently follow attacker-controlled profile information or invoke attacker-controlled endpoints. ### Impact Assessmen ...[truncated 496 chars]
Remediation
## Remediation Suggestions 1. Replace address-only authentication with cryptographic proof of wallet ownership. 2. Have the server issue a unique, unpredictable, short-lived nonce for each authentication or update attempt. 3. Require the wallet to sign a structured message containing at least: - The nonce - The agent ID - The HTTP method and intended operation - A canonical representation or digest of the changed fields - The target domain and chain ID - An issuance time and expiration time 4. Verify the signature against the registered wallet address before accepting any modification. 5. Mark each nonce as consumed after successful verification and reject reused, expired, or mismatched nonces. 6. Consider established wallet-authentication standards, such as Sign-In with Ethereum, with appropriate domain separation and replay protection. 7. Apply rate limiting, update auditing, owner notifications, and recovery procedures for unauthorized profile changes. 8. Update `SKILL.md` so it does not describe a public wallet address as authentication and clearly documents the required signing flow.
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (10)

External Transmission

Medium
Category
Data Exfiltration
Content
## Register Your Agent

```bash
POST https://api.deepbluebase.xyz/agents/register
Content-Type: application/json

{
Confidence
50% confidence
Finding
Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

External Transmission

Medium
Category
Data Exfiltration
Content
## Register Your Agent

```bash
POST https://api.deepbluebase.xyz/agents/register
Content-Type: application/json

{
Confidence
50% confidence
Finding
Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

External Transmission

Medium
Category
Data Exfiltration
Content
## Register Your Agent

```bash
POST https://api.deepbluebase.xyz/agents/register
Content-Type: application/json

{
Confidence
50% confidence
Finding
Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

External Transmission

Medium
Category
Data Exfiltration
Content
## Register Your Agent

```bash
POST https://api.deepbluebase.xyz/agents/register
Content-Type: application/json

{
Confidence
50% confidence
Finding
Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

External Transmission

Medium
Category
Data Exfiltration
Content
## Register Your Agent

```bash
POST https://api.deepbluebase.xyz/agents/register
Content-Type: application/json

{
Confidence
50% confidence
Finding
Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

External Transmission

Medium
Category
Data Exfiltration
Content
## Register Your Agent

```bash
POST https://api.deepbluebase.xyz/agents/register
Content-Type: application/json

{
Confidence
50% confidence
Finding
Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

External Transmission

Medium
Category
Data Exfiltration
Content
## Register Your Agent

```bash
POST https://api.deepbluebase.xyz/agents/register
Content-Type: application/json

{
Confidence
50% confidence
Finding
Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

Missing User Warnings

Medium
Confidence
96% confidence
Finding
The skill states only in a brief note that providing an X handle causes the service to automatically fetch and publicly use the associated X profile photo as the agent avatar. This is a privacy and consent issue because users may disclose a social identifier expecting only profile linking, not automatic retrieval and republication of profile imagery on a public directory.

External Transmission

Medium
Category
Data Exfiltration
Content
## Proxy-Call Any Listed Endpoint

```bash
POST https://api.deepbluebase.xyz/bluepages/call
Content-Type: application/json
# Payment: $0.002 USDC via x402 (DeepBlue facilitator fee)
Confidence
85% confidence
Finding
The proxy-call feature allows the service to make arbitrary upstream calls on the user's behalf based on a supplied URL and method. In a skill context, this can enable unintended third-party data disclosure, interaction with untrusted endpoints, or use as a network relay without clear restrictions, making it more dangerous than ordinary fixed-destination API calls.

External Transmission

Medium
Category
Data Exfiltration
Content
# Payment: $0.002 USDC via x402 (DeepBlue facilitator fee)

{
  "url": "https://api.aixbt.tech/v2/signals",
  "method": "GET",
  "params": { "limit": 5 }
}
Confidence
50% confidence
Finding
Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

Static analysis

No suspicious patterns detected.