ClawHub
Back to skill

Security audit

performance-engineer

Security checks for vulnerabilities and agentic risk

Overview

This performance skill is not malware, but it needs review because some examples can permanently change databases or Linux hosts without enough safety guidance.

Review the examples before installing or using this skill on real systems. Treat the SQL and Linux tuning snippets as illustrative only unless an operator validates them in a non-production environment, confirms compatibility, backs up existing settings, and has a tested rollback plan.

A.I.G

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
references/examples.md:354
Finding

Unguarded Privileged System-Wide Performance Tuning

Content
View full analysis
$cpu done # Disable CPU frequency scaling echo 0 > /sys/devices/system/cpu/intel_pstate/no_turbo # 2. Memory optimization # Transparent Huge Pages echo never > /sys/kernel/mm/transparent_hugepage/enabled echo never > /sys/kernel/mm/transparent_hugepage/defrag # Swappiness (prefer RAM over swap) echo 10 > /proc/sys/vm/swappiness # 3. Network optimization cat >> /etc/sysctl.conf << EOF # Network performance tuning net.core.rmem_max = 134217728 net.core.wmem_max = 134217728 net.ipv4.tcp_rmem = 4096 87380 134217728 net.ipv4.tcp_wmem = 4096 65536 134217728 net.core.netdev_max_backlog = 5000 net.ipv4.tcp_congestion_control = bbr net.ipv4.tcp_notsent_lowat = 16384 net.ipv4.tcp_tw_reuse = 1 net.ipv4.tcp_fin_timeout = 15 net.ipv4.tcp_keepalive_time = 300 net.ipv4.tcp_keepalive_probes = 5 net.ipv4.tcp_keepalive_intvl = 15 EOF sysctl -p # 4. Disk I/O optimization # Set scheduler for SSDs for disk in /sys/block/sd*/queue/scheduler; do echo noop > $disk done # Increase read-ahead for disk in /sys/block/sd*/queue/read_ahead_kb; do echo 256 > $disk done # 5. File system tuning # Increase file descriptors echo "* soft nofile 65535" >> /etc/security/limits.conf echo "* hard nofile 65535" >> /etc/security/limits.conf ``` ### Technical Analysis The example makes direct, unguarded changes to active kernel controls and persistent host-wide configuration. It does not: - Verify the operating system, kernel version, hardware, workload, or virtualization environment. - Confirm that the BBR congestion-control algorithm or the `noop` disk scheduler is supported. - Distinguish SSDs from rotational disks despite claiming to configure ...[truncated 2884 chars]
Remediation
View remediation

SkillSpector

Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (4)

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The manifest description uses broad phrasing such as "Use when: performance analysis framework, application profiling techniques, load testing strategies," which reads more like topical keywords than explicit activation conditions. It does not clearly specify what user requests should or should not invoke the skill, increasing the risk of unintended activation for general performance-related conversation.

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 · references/examples.md (reported line 197)May include surrounding context.

md
};

export default function() {
    let response = http.get('https://api.example.com/endpoint');
    
    check(response, {
        'status is 200': (r) => r.status === 200,

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The SQL examples contain operationally significant commands such as CREATE INDEX, CREATE MATERIALIZED VIEW, REFRESH MATERIALIZED VIEW, ALTER SYSTEM, and partition creation, but they are presented without warnings about locking, resource usage, persistence, or production impact. Readers could apply them blindly to live databases and trigger outages, performance regressions, or hard-to-reverse configuration changes.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

This example includes host-level tuning commands that directly modify kernel, CPU, disk, limits, and JVM settings without any safety warnings, scoping guidance, rollback steps, or environment restrictions. In a skill intended as reusable guidance, users may copy and run the script on production or incompatible systems, causing instability, degraded security posture, or service disruption.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.