ClawHub
Back to skill

Security audit

Error Recovery

Security checks across malware telemetry and agentic risk

Overview

This is a legitimate error-recovery skill, but some automatic recovery steps can press keys, delete lock files, and persist session content without enough scoping or safeguards.

Review before installing. Do not run the tmux rate-limit script from cron as-is; restrict it to known agent panes and add the documented second safety check before sending Enter. Use dry-run or confirmation for lock cleanup and rollback, and enable transcript scavenging only with explicit opt-in, redaction, and retention limits.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • 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
Findings (3)

Missing User Warnings

Medium
Confidence
87% confidence
Finding
The skill recommends automatic cleanup and state-changing recovery actions such as clearing locks and modifying recovery state without an explicit warning, confirmation gate, or dry-run mode. In a crash-recovery context, that can cause unintended deletion or mutation of files/state after a false positive, potentially destroying evidence, corrupting concurrent work, or rolling forward from an unsafe assumption.

Missing User Warnings

Medium
Confidence
90% confidence
Finding
The document recommends `truncate conversation head` as a last-resort recovery step and explicitly notes that it is irreversible and loses early context, but it does not require a user warning, confirmation, or other safety gate before destructive context loss. In an error-recovery skill, this is more dangerous because the behavior may be triggered under failure conditions where the agent is already unstable, increasing the chance of silent loss of instructions, commitments, or safety-relevant context.

Missing User Warnings

Medium
Confidence
96% confidence
Finding
The document explicitly instructs a daemon to analyze dead-session transcripts, including thinking blocks, and archive extracted findings without any privacy, minimization, or secret-handling safeguards. In an agent error-recovery skill, this is more dangerous because transcripts commonly contain credentials, personal data, internal reasoning, and sensitive tool outputs, so automatic scavenging can create a durable secondary leak surface.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.