Agi Terminal Helper
PassAudited by ClawScan on May 1, 2026.
Overview
This instruction-only skill is a safety runbook for terminal execution and is aligned with its stated purpose, though users should remember that local exec commands are powerful.
This appears safe to install if you want a runbook for using OpenClaw exec. Expect it to help draft and explain terminal commands, and only approve commands that match your intent, especially anything that writes files, installs software, touches secrets, uses sudo, or changes system settings.
Findings (2)
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
The agent may propose terminal commands in your workspace; those commands can affect local files or system state if you approve risky actions.
The skill gives the agent a workflow for using local shell execution, which can read or modify local files if misused. The same instructions add review, sandboxing, read-only defaults, and confirmation requirements, so this is purpose-aligned rather than suspicious.
Before calling `exec`, I will say ... what files it might read/write ... Default to read-only exploration ... Explicit confirmation for anything risky
Keep the stated confirmation process in place, especially for writes, installs, secrets, privileged commands, and long-running processes.
You have less publisher/source context than with a skill that links to a public repository or homepage.
The registry entry provides limited provenance information. Because the skill is instruction-only, has no install spec, and includes no code files, this is a low-impact provenance note rather than a security concern.
Source: unknown; Homepage: none
Review the SKILL.md text before installing and prefer skills with clear source provenance when possible.