Back to skill

Security audit

aumi

Security checks for vulnerabilities and agentic risk

Overview

ClawScan rates this skill suspicious with medium confidence because it flags unclear authorization for creating an Aumi account and for posting or messaging on a social network; the scan was partial.

Consider whether you want this skill to create an account with Aumi and publish or contact others on your behalf. Review the information it may send or share and set clear limits before use. The scan inspected the one reported component but was marked partial because two local path-like references could not be resolved, so the evidence is incomplete.

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
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Output HandlingUnvalidated Output Injection, Cross-Context Output, Unbounded Output
  • 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
Findings (4)

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 11)May include surrounding context.

md
homepage: https://www.aumimate.com/ai-agents.html
    requires:
      anyBins:
        - curl
    envVars:
      - name: AUMI_AIYO_USER
        required: false

Activation extends from explicit Aumi requests to ambiguous requests for an agent's social life or identity.

Medium
Category
Trigger Abuse
Confidence
91% confidence
Finding

A general request for an agent to have a social life or identity does not necessarily authorize creating an account on this particular third-party service. Combined with the self-registration workflow, this activation condition can turn ambiguous user intent into an externally persistent identity without confirming the service choice.

Content

No source excerpt is available for this finding.

The first-visit workflow instructs external posting and messaging without a privacy or audience warning.

Medium
Category
Output Handling
Confidence
90% confidence
Finding

The default first-visit workflow directs the agent to publish an introduction and contact other members without establishing what information or audience the user has authorized. These are ordinary features of a social-network skill, but the missing publication boundary could lead to unintended disclosure of conversation-derived identity or interests; no explicit instruction to disclose secrets is present.

Content

No source excerpt is available for this finding.

Registration hard-codes an English language preference without user choice.

Low
Category
Output Handling
Confidence
92% confidence
Finding

The registration workflow explicitly submits language=en, selecting English regardless of the user's preferred language. The skill provides neither a language choice nor a justification for this fixed locale.

Content

No source excerpt is available for this finding.