ClawHub
Back to skill

Security audit

ChatGPT Import

Security checks for vulnerabilities and agentic risk

Overview

This skill does what it says: imports ChatGPT history into OpenClaw search, while requiring care because chat contents are stored locally and sent to OpenAI for embeddings.

Install only if you are comfortable importing your ChatGPT history into local markdown and SQLite files and sending conversation chunks to OpenAI for embeddings. Review or filter chats that may contain secrets, personal data, business data, or regulated information, and use a scoped OpenAI API key where possible.

A.I.G

Vulnerability Patterns
  • 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
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs

SkillSpector

Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Rogue AgentSelf-Modification, Session Persistence
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (6)

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
95% confidence
Finding

The skill instructs the user to run local scripts that read exported conversation files, write a SQLite database, access environment variables for an API key, and send conversation contents over the network to OpenAI for embedding, but it declares no explicit tool scope or permissions. This creates a security transparency and least-privilege problem: operators cannot easily tell from the manifest that the skill handles sensitive local data and performs external data exfiltration to a third-party API.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

The guide tells users to export and handle a full ChatGPT account archive, explicitly noting it contains all conversations, but it does not warn that this archive may include sensitive personal, business, authentication, or regulated data. In the context of an import skill, that omission increases the chance users will move, store, or process highly sensitive data without adequate review, minimization, or secure handling.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The script sends raw markdown conversation content to the OpenAI embeddings API, which is an external third party, but it does not provide an explicit warning, consent prompt, or redaction step before transmission. In the context of importing ChatGPT history, those files are likely to contain sensitive personal, business, or credential-bearing data, so silent export materially increases privacy and compliance risk.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
98% confidence
Finding

This code performs deliberate external transmission of the imported markdown text to https://api.openai.com/v1/embeddings. Because the skill's purpose is to ingest historical conversations into searchable memory, the transmitted data may include highly sensitive chat history, making the external network call security-relevant even if it is functionally intended.

Content

Scanner excerpt · scripts/bulk_embed.py (reported line 70)May include surrounding context.

python
"""Call OpenAI embeddings API."""
    data = json.dumps({"input": texts, "model": model}).encode()
    req = urllib.request.Request(
        "https://api.openai.com/v1/embeddings", data=data,
        headers={"Authorization": f"Bearer {api_key}", "Content-Type": "application/json"},
    )
    for attempt in range(3):

Session Persistence

Medium
Category
Rogue Agent
Confidence
60% confidence
Finding

Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.

Content

Scanner excerpt · scripts/bulk_embed.py (reported line 148)May include surrounding context.

python
print(f"Embedding done in {time.time() - start_time:.1f}s")

    # Write DB
    db_path = args.db
    if os.path.exists(db_path):
        print(f"Error: output DB already exists at {db_path}", file=sys.stderr)

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

This code writes extracted ChatGPT conversation text into local markdown files, which can include sensitive personal or proprietary content. Although the script prints progress messages, there is no user-facing warning, comment, or docstring disclosure about persisting conversation data to disk.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.