ClawHub
Back to skill

Security audit

huawei-cloud-cloudrobo-train-metrics

Security checks for vulnerabilities and agentic risk

Overview

The skill adds clearly disclosed training-metric logging, with no evidence of hidden access or unrelated activity in its bundled files.

Confirm the training project and metric directory, use a separate output directory for each run to preserve previous metadata, and grant no cloud credentials. Obtain the optional ocms wheel from a verified provider source; that external dependency was unavailable for inspection. The platform may collect metrics written to METRICS, as disclosed.

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
  • Behavioral ASTexec() Call, eval() Call, Dynamic Import
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (2)

Skill declares no tool scope ('permissions' or 'allowed-tools') but code capabilities were detected: env, file_read.

Medium
Category
MCP Least Privilege
Confidence
70% confidence
Finding

Without declared permissions the skill's intent is opaque and cannot be validated.

Content

No source excerpt is available for this finding.

Dynamic attribute access via getattr()

Low
Category
Dangerous Code Execution
Confidence
50% confidence
Finding

Dynamic getattr() with a non-literal attribute name can access arbitrary object attributes, potentially bypassing access controls.

Content

Scanner excerpt · scripts/ocms_recorder.py (reported line 207)May include surrounding context.

python
break
                if not all_present or not values:
                    continue
                getattr(self._summary, chart_name).add(step, values)
        except Exception as exc:
            logger.warning(f"ocms recording failed: {exc}")
            self.enabled = False