Back to skill

Security audit

Ainative Agent Framework

Security checks for vulnerabilities and agentic risk

Overview

The skill is not destructive, but its examples send conversation data to AINative and can persist user-derived content in cloud memory without clear consent or retention controls.

Install only if you intend to use AINative-hosted services and are comfortable sending prompts, conversation history, memory contents, recall queries, and feedback to that provider. Do not use the examples with secrets, regulated data, customer data, proprietary code, or security findings unless your organization has approved that data flow and you have a retention/deletion plan.

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:73
Finding

Unbounded Transmission and Persistence of Potentially Sensitive Conversation Data

Content
View full analysis
str: self.messages.append({"role": "user", "content": user_input}) resp = requests.post( "https://api.ainative.studio/v1/public/chat/completions", headers={"X-API-Key": self.api_key}, json={ "model": "claude-3-5-sonnet-20241022", "messages": [ {"role": "system", "content": self.system_prompt}, *self.messages ], "max_tokens": 2048, } ).json() reply = resp["choices"][0]["message"]["content"] self.messages.append({"role": "assistant", "content": reply}) return reply def remember(self, fact: str): """Persist something to ZeroMemory.""" requests.post( "https://api.ainative.studio/api/v1/public/memory/v2/remember", headers={"X-API-Key": self.api_key}, json={"content": fact, "memory_type": "episodic"} ) def recall(self, query: str) -> list: """Retrieve relevant memories.""" resp = requests.post( "https://api.ainative.studio/api/v1/public/memory/v2/recall", headers={"X-API-Key": self.api_key}, json={"query": query, "limit": 5} ).json() return [m["content"] for m in resp.get("memories", [])] # Usage agent = AINativeAgent("ak_your_key", "You are a helpful coding assistant.") response = agent.think("How do I implement rate limiting in FastAPI?") agent.remember(f"User as ...[truncated 3762 chars]
Remediation
View remediation
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 (14)

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The example agent sends user input and accumulated conversation history to an external API endpoint but the skill does not clearly disclose that prompts and prior messages leave the local environment. This creates a privacy and data-handling risk, especially if operators paste secrets, proprietary code, or personal data into the agent.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
91% confidence
Finding

This duplicate finding points to the same chat completion request that exports user input and conversation context to a remote service. The core issue is undisclosed external transmission of potentially sensitive prompts and history.

Content

Scanner excerpt · SKILL.md (reported line 78)May include surrounding context.

md
def think(self, user_input: str) -> str:
        self.messages.append({"role": "user", "content": user_input})

        resp = requests.post(
            "https://api.ainative.studio/v1/public/chat/completions",
            headers={"X-API-Key": self.api_key},
            json={

External Transmission

Medium
Category
Data Exfiltration
Confidence
91% confidence
Finding

This duplicate finding points to the same chat completion request that exports user input and conversation context to a remote service. The core issue is undisclosed external transmission of potentially sensitive prompts and history.

Content

Scanner excerpt · SKILL.md (reported line 78)May include surrounding context.

md
def think(self, user_input: str) -> str:
        self.messages.append({"role": "user", "content": user_input})

        resp = requests.post(
            "https://api.ainative.studio/v1/public/chat/completions",
            headers={"X-API-Key": self.api_key},
            json={

External Transmission

Medium
Category
Data Exfiltration
Confidence
90% confidence
Finding

The hardcoded external API domain in the chat example confirms that model inputs are sent off-box to a hosted service. In a skill file, this becomes a practical privacy risk because the example lacks any notice about external data handling.

Content

Scanner excerpt · SKILL.md (reported line 79)May include surrounding context.

md
self.messages.append({"role": "user", "content": user_input})

        resp = requests.post(
            "https://api.ainative.studio/v1/public/chat/completions",
            headers={"X-API-Key": self.api_key},
            json={
                "model": "claude-3-5-sonnet-20241022",

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The documentation explicitly demonstrates persisting user-derived content to a remote memory service, including a usage example that stores part of a model response derived from a user request. Without a clear warning, consent flow, or data-minimization guidance, users may unknowingly transmit sensitive prompts or outputs to external storage.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
94% confidence
Finding

This duplicate finding refers to the same remote memory persistence call. Storing arbitrary content remotely can create confidentiality and compliance issues when examples do not instruct users to limit or sanitize stored data.

Content

Scanner excerpt · SKILL.md (reported line 97)May include surrounding context.

md
def remember(self, fact: str):
        """Persist something to ZeroMemory."""
        requests.post(
            "https://api.ainative.studio/api/v1/public/memory/v2/remember",
            headers={"X-API-Key": self.api_key},
            json={"content": fact, "memory_type": "episodic"}

External Transmission

Medium
Category
Data Exfiltration
Confidence
94% confidence
Finding

This duplicate finding refers to the same remote memory persistence call. Storing arbitrary content remotely can create confidentiality and compliance issues when examples do not instruct users to limit or sanitize stored data.

Content

Scanner excerpt · SKILL.md (reported line 97)May include surrounding context.

md
def remember(self, fact: str):
        """Persist something to ZeroMemory."""
        requests.post(
            "https://api.ainative.studio/api/v1/public/memory/v2/remember",
            headers={"X-API-Key": self.api_key},
            json={"content": fact, "memory_type": "episodic"}

External Transmission

Medium
Category
Data Exfiltration
Confidence
93% confidence
Finding

The memory endpoint URL shows that remembered content is persisted to an external hosted service rather than remaining local. This is dangerous when paired with examples that encourage storing user-derived facts without consent or sensitivity checks.

Content

Scanner excerpt · SKILL.md (reported line 98)May include surrounding context.

md
def remember(self, fact: str):
        """Persist something to ZeroMemory."""
        requests.post(
            "https://api.ainative.studio/api/v1/public/memory/v2/remember",
            headers={"X-API-Key": self.api_key},
            json={"content": fact, "memory_type": "episodic"}
        )

External Transmission

Medium
Category
Data Exfiltration
Confidence
88% confidence
Finding

This duplicate finding refers to the same remote memory recall API call. Even a query string can leak confidential project names, incident details, or user identifiers to an external service.

Content

Scanner excerpt · SKILL.md (reported line 105)May include surrounding context.

md
def recall(self, query: str) -> list:
        """Retrieve relevant memories."""
        resp = requests.post(
            "https://api.ainative.studio/api/v1/public/memory/v2/recall",
            headers={"X-API-Key": self.api_key},
            json={"query": query, "limit": 5}

External Transmission

Medium
Category
Data Exfiltration
Confidence
88% confidence
Finding

This duplicate finding refers to the same remote memory recall API call. Even a query string can leak confidential project names, incident details, or user identifiers to an external service.

Content

Scanner excerpt · SKILL.md (reported line 105)May include surrounding context.

md
def recall(self, query: str) -> list:
        """Retrieve relevant memories."""
        resp = requests.post(
            "https://api.ainative.studio/api/v1/public/memory/v2/recall",
            headers={"X-API-Key": self.api_key},
            json={"query": query, "limit": 5}

External Transmission

Medium
Category
Data Exfiltration
Confidence
87% confidence
Finding

The recall endpoint URL indicates that memory searches are performed against an external service, exposing query content and usage patterns. The absence of disclosure makes this more risky from a privacy and governance perspective.

Content

Scanner excerpt · SKILL.md (reported line 106)May include surrounding context.

md
def recall(self, query: str) -> list:
        """Retrieve relevant memories."""
        resp = requests.post(
            "https://api.ainative.studio/api/v1/public/memory/v2/recall",
            headers={"X-API-Key": self.api_key},
            json={"query": query, "limit": 5}
        ).json()

External Transmission

Medium
Category
Data Exfiltration
Confidence
86% confidence
Finding

This duplicate finding refers to the same RLHF feedback submission endpoint. Feedback content and session metadata may expose sensitive security or product information if copied directly from real interactions.

Content

Scanner excerpt · SKILL.md (reported line 157)May include surrounding context.

text

```python
requests.post(
    "https://api.ainative.studio/api/v1/public/zerodb/rlhf/feedback",
    headers={"X-API-Key": "ak_your_key"},
    json={

External Transmission

Medium
Category
Data Exfiltration
Confidence
86% confidence
Finding

This duplicate finding refers to the same RLHF feedback submission endpoint. Feedback content and session metadata may expose sensitive security or product information if copied directly from real interactions.

Content

Scanner excerpt · SKILL.md (reported line 157)May include surrounding context.

text

```python
requests.post(
    "https://api.ainative.studio/api/v1/public/zerodb/rlhf/feedback",
    headers={"X-API-Key": "ak_your_key"},
    json={

External Transmission

Medium
Category
Data Exfiltration
Confidence
85% confidence
Finding

The RLHF endpoint URL confirms export of evaluation data to an external service. If operators include real session IDs, bug details, or security observations, that information may leave the trusted environment without clear notice.

Content

Scanner excerpt · SKILL.md (reported line 158)May include surrounding context.

python
requests.post(
    "https://api.ainative.studio/api/v1/public/zerodb/rlhf/feedback",
    headers={"X-API-Key": "ak_your_key"},
    json={
        "session_id": "sess-123",