Skillv1.0.0

ClawScan security

Boot Kpi Analyzer · ClawHub's context-aware review of the artifact, metadata, and declared behavior.

Scanner verdict

BenignMar 6, 2026, 8:13 AM

Verdict: benign
Confidence: high
Model: gpt-5-mini
Summary: The skill's code and runtime instructions align with its stated purpose (local analysis of boot/shutdown KPI logs); it reads specific local log directories and may curl a private IP but requests no unrelated credentials or installs.
Guidance: This skill appears to do exactly what it claims: parse log files under /home/lixiang/... and optionally query a private web API. Before installing or running it: (1) confirm the indicated directories exist and that you want the agent to read them (the paths are hardcoded to /home/lixiang/...); (2) confirm you trust the private IP endpoint (http://10.122.86.46:9999) the skill may curl; (3) review the included script if you need to be certain it doesn't transmit data elsewhere — it appears to only parse local logs; (4) be aware the agent will run python3 and shell curl commands, so run in a principle-of-least-privilege context or on a machine/account that is allowed to expose those logs. If any of these assumptions don't hold (different home path, untrusted network endpoint), do not enable/auto-run the skill until modified.

Review Dimensions

Purpose & Capability: okName/description describe analysis of boot/shutdown KPI data; the included script and SKILL.md operate on the exact log directories and files described (boot_kpi.log, unit_kpi.log, shutdown-monitor-units.log, report.json). No unrelated services, credentials, or broad system access are requested.
Instruction Scope: noteInstructions explicitly require reading files under /home/lixiang/... and optionally curling a private IP (http://10.122.86.46:9999). This is consistent with the skill's purpose, but the instructions hardcode a specific user path and instruct the agent to run shell commands (python3 and curl), which means the agent will access local files and make network requests when invoked.
Install Mechanism: okNo install spec; the skill is instruction+script only. Nothing is downloaded or written to disk at install time beyond the normal skill files, which is the lowest-risk pattern.
Credentials: okThe skill declares no required environment variables or credentials. Its file- and network-access needs (local log dirs and a private web API) are proportionate to the stated log-analysis purpose.
Persistence & Privilege: notealways:false (default) and model invocation is enabled (default). Autonomous invocation is the platform default; since the skill reads local files and can run curl/python3, users should be aware an agent could execute these actions autonomously if allowed. There is no always:true or other elevated persistence requested.