ClawHub
Back to skill

Security audit

Agent Archive

Security checks for vulnerabilities and agentic risk

Overview

This skill appears to be a user-invoked conversation export tool that writes local archives, which is sensitive but matches its stated purpose.

Install only if you want local archives of full agent sessions. Before sharing, committing, backing up, or indexing exported files, review them for secrets, personal data, file paths, proprietary code, and tool outputs; prefer a narrower export mode or manual redaction for sensitive sessions.

SkillSpector

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 (4)

Missing User Warnings

Medium
Confidence
91% confidence
Finding
The skill is explicitly designed to export and persist conversations to disk, but it does not warn that archives may contain sensitive user data, secrets, filesystem context, or tool outputs. Users may invoke it expecting convenience, not realizing it creates durable local copies of potentially confidential material.

Missing User Warnings

Medium
Confidence
94% confidence
Finding
The replay/full export guidance says to preserve tool arguments, results, timing, errors, and cwd-related metadata, but omits any warning that these fields may include secrets, tokens, file paths, proprietary code, or other sensitive operational context. This increases the risk of unintentional data retention and later disclosure through archived logs.

Missing User Warnings

Medium
Confidence
90% confidence
Finding
This script exports entire conversations, including raw message content and tool results, directly to disk with no redaction, sensitivity checks, or explicit warning to the operator. In the context of agent sessions, tool results commonly contain secrets, filesystem contents, tokens, API responses, or other sensitive material, so exporting them verbatim can create persistent local data exposure and accidental leakage through backups, commits, or sharing.

Ssd 3

Medium
Confidence
95% confidence
Finding
The skill instructs the agent to preserve full conversational and tool execution history, including user input, tool arguments, tool results, timing, errors, and replay structure. In this context, that creates a real data leakage risk because archives become a consolidated record of sensitive prompts, outputs, credentials, local paths, and other artifacts that may later be accessed, copied, indexed for RAG, or committed inadvertently.

Static analysis

No suspicious patterns detected.