ClawHub
Back to skill

Security audit

Mobile EU Compliance

Security checks for vulnerabilities and agentic risk

Overview

This skill reviews mobile-app compliance and produces a report without changing code unless asked.

Limit assessment to your intended app project and treat the output as engineering guidance. Have your compliance owner verify current legal sources; the referenced docs/SOURCES.md is not bundled. Manual inspection found no support for the scanner's behavioral concerns, and the supplied static scan reported no findings.

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
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Memory PoisoningPersistent Context Injection, Context Window Stuffing, Memory Manipulation
  • Tool MisuseTool Parameter Abuse, Chaining Abuse, Unsafe Defaults
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Findings (4)

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
80% confidence
Finding

Tool parameters are crafted to achieve unintended or unsafe behavior. Parameter abuse can bypass intended safety checks (e.g. shell=True, --force, dangerous glob patterns).

Content

Scanner excerpt · references/evidence-templates.md (reported line 38)May include surrounding context.

| Pairing / onboarding | app + device | OnboardingFlow.kt, firmware spec section 4 | | App update enforcement | app + backend | UpdateGate.kt, remote config min_version_code | | Firmware update | app + device + backend | DfuManager.kt, bootloader signature check | | Deletion | app + backend + device | AccountDeletion.kt, API DELETE /me, reset command 0x7F |

text

## 2. Evidence rules

Memory Manipulation

High
Category
Memory Poisoning
Confidence
80% confidence
Finding

Skill manipulates agent memory, state, or stored context. Memory corruption can alter personality, override safety rules, or cause unpredictable behavior.

Content

Scanner excerpt · references/evidence-templates.md (reported line 46)May include surrounding context.

md
| Status | Needs |
|---|---|
| Met | A file and line, a config, a test, a CI job, a published URL or a named process document |
| Partial | Evidence for part of the requirement and a clear statement of what is missing |
| Missing | Nothing found. Say where the agent looked |
| Not applicable | The reason, for example "no payment function, EN 18031-3 does not apply" |

Autonomous Decision Making

Medium
Category
Excessive Agency
Confidence
75% confidence
Finding

Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.

Content

Scanner excerpt · SKILL.md (reported line 168)May include surrounding context.

md
// BAD: analytics and ad SDKs start in Application.onCreate for every user
override fun onCreate() { super.onCreate(); Analytics.init(this); Ads.init(this) }

// GOOD: start only after the user agreed; strictly necessary storage (session, settings) needs no consent
fun onConsentChanged(consent: Consent) {
    if (consent.analytics) Analytics.init(context) else Analytics.disable(context)
}

Autonomous Decision Making

Medium
Category
Excessive Agency
Confidence
75% confidence
Finding

Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.

Content

Scanner excerpt · references/gdpr.md (reported line 47)May include surrounding context.

md
For apps this covers more than cookies:
- Reading the advertising ID, writing analytics IDs, SDK caches used for tracking, device fingerprinting.
- Strictly necessary (no consent needed): login session, user settings, offline data of the feature the user uses.
- Consent must be collected before the SDK starts, not after.

## Third-party SDKs