ClawHub
Back to skill

Security audit

Browser Zombie Cleaner

Security checks for vulnerabilities and agentic risk

Overview

The skill is a disclosed browser-cleanup tool, but its kill mode can be widened enough to close unrelated browser processes.

Install only if you are comfortable with a local shell script that can terminate browser processes. Use the default detect mode first, avoid overriding --pattern, and use --kill only when the listed processes are clearly OpenClaw browser leftovers. There is no evidence of credential access, exfiltration, persistence, or remote code loading, but the cleanup controls should be tightened before unattended use.

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

Error
Location
scripts/cleanup-zombie-browsers.sh:38
Finding

User-Controlled Filters Can Broaden Destructive Browser Process Termination

Content
View full analysis
/dev/null)" || continue if [[ "$verify_cmdline" != *"$OPENCLAW_BROWSER_PATTERN"* ]]; then log "SKIP PID $pid: re-verification failed (cmdline changed)" continue fi fi # Count child processes that will also be terminated child_count=0 if [[ "$OS" == "linux" ]]; then child_count=$(grep -r "PPid:.*$pid" /proc ...[truncated 3832 chars]
Remediation
View remediation
&2 exit 2 } [[ "$GRACE_PERIOD_SECONDS" =~ ^[0-9]+$ ]] || { echo "Invalid --grace value" >&2 exit 2 } ``` Destructive mode should enforce a reasonable nonzero minimum age unless an explicitly named unsafe override is separately confirmed. 4. **Revalidate every safety property immediately before signaling.** Confirm current ownership, executable identity, canonical user-data directory, PPID/orphan state, minimum age, and process start time. 5. **Defend against PID reuse.** Record the process start time during discovery and compare it with the start time immediately before each signal. 6. **Avoid signaling unchecked process groups.** Signal only the verified candidate PID and descendants that have been independently enumerated and validated. If process-group signaling remains necessary, verify every group member first and abort when an unrelated member is present. 7. **Apply the same checks before SIGKILL.** Repeat identity and ownership validation after the grace period so that a changed or reused PID is not forcefully terminated. 8. **Update documentation and help text.** Clearly state the risks of any safety-control overrides, or remove `--pattern` from the public interface entirely. ]]>

SkillSpector

Vulnerability Patterns
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep