Security checks for vulnerabilities and agentic risk
Overview
This skill has a coherent memory purpose, but it can automatically send and retain user context with an external service and recommends an unpinned MCP install, so users should review it before installing.
Install only if you are comfortable sending memory contents, search queries, task summaries, and preferences to Lithtrix for cross-session persistence. Use a pinned and reviewed MCP package version, limit the API key's scope where possible, avoid storing secrets or sensitive personal, financial, legal, health, proprietary, or credential data, and ensure users can approve, review, and delete stored memories.
The prompt instructions tell the agent to automatically store summaries after meaningful tasks or when learning a user preference, but provide no boundary on sensitive data. This can lead to silent persistence of personal data, confidential task details, or regulated information across sessions in a remote system, increasing privacy and compliance risk.
Content
No source excerpt is available for this finding.
Missing User Warnings
Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding
The skill is explicitly designed for persistent cross-session memory and says the agent can store and retrieve memories across sessions, but it does not warn about privacy, retention, consent, or what categories of data should never be stored. In practice this encourages operators to externalize potentially sensitive conversation history and preferences to a third-party service without clear safeguards.
Content
No source excerpt is available for this finding.
External Transmission
Medium
Category
Data Exfiltration
Confidence
50% 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 27)May include surrounding context.
Get an API key. Register your agent at lithtrix.ai or call the API directly:
text
POST https://api.lithtrix.ai/v1/register
Body: { "agent_name": "your-agent-name" }
The response contains your ltx_... key.
External Transmission
Medium
Category
Data Exfiltration
Confidence
50% 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 68)May include surrounding context.
Get an API key. Register your agent at lithtrix.ai or call the API directly:
text
POST https://api.lithtrix.ai/v1/register
Body: { "agent_name": "your-agent-name" }
The response contains your ltx_... key.
External Transmission
Medium
Category
Data Exfiltration
Confidence
50% 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 96)May include surrounding context.
Get an API key. Register your agent at lithtrix.ai or call the API directly:
text
POST https://api.lithtrix.ai/v1/register
Body: { "agent_name": "your-agent-name" }
The response contains your ltx_... key.
External Transmission
Medium
Category
Data Exfiltration
Confidence
50% 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 120)May include surrounding context.
Get an API key. Register your agent at lithtrix.ai or call the API directly:
text
POST https://api.lithtrix.ai/v1/register
Body: { "agent_name": "your-agent-name" }
The response contains your ltx_... key.
Rp1
Medium
Category
MCP Rug Pull
Confidence
97% confidence
Finding
The skill instructs users to run an MCP server via npx lithtrix-mcp without pinning a specific version or integrity-checked artifact. That creates a supply-chain risk: future package updates or a compromised publisher account could cause arbitrary code to be fetched and executed in the agent environment, which is especially sensitive because the skill also relies on an API key.
Content
No source excerpt is available for this finding.
External Transmission
Medium
Category
Data Exfiltration
Confidence
94% confidence
Finding
This endpoint is used to transmit arbitrary memory values to a third-party API, and the skill's surrounding guidance encourages storing user preferences and task summaries without clear limits. In context, the dangerous part is not merely the outbound request but the unsafeguarded exfiltration and retention of potentially sensitive conversation-derived data.
Content
Scanner excerpt · SKILL.md (reported line 58)May include surrounding context.
Store a memory
text
POST https://api.lithtrix.ai/v1/memory
Authorization: Bearer $LITHTRIX_API_KEY
{
"key": "user_preference_theme",
External Transmission
Medium
Category
Data Exfiltration
Confidence
90% confidence
Finding
The semantic search endpoint sends user-derived queries and likely causes related stored memories to be processed by an external service. Given the instruction to search past context before answering, this can expose private user history or sensitive topics to a third party without an explicit warning or opt-in.
Content
Scanner excerpt · SKILL.md (reported line 74)May include surrounding context.
Semantic search over memory
text
POST https://api.lithtrix.ai/v1/memory/search
Authorization: Bearer $LITHTRIX_API_KEY
{
"query": "what did the user ask about last week?",
Natural-Language Policy Violations
Medium
Category
Not specified by scanner
Confidence
88% confidence
Finding
The example normalizes using stored memory to change future behavior automatically, which can undermine user expectations and consent at response time. While a language preference is low sensitivity by itself, the pattern encourages agents to act on historical data without transparency, and the same mechanism could be applied to more sensitive preferences or prior context.
