Pre-Task Checklist

Mandatory pre-task verification protocol to prevent forgetfulness, stale data errors, and context loss. Use when: (1) starting any non-trivial task, (2) befo...

This skill appears coherent and implements a reasonable pre-task checklist, but it will read files from your workspace and may print their contents and attempt SSH connections using whatever SSH keys/config are available. Before loading or sourcing the scripts: (1) review scripts/checklist.sh and scripts/verify.py yourself (they are included) to ensure you’re comfortable with which files they read; (2) be aware that outputs from running the skill could expose sensitive file contents or tokens if those files exist; (3) if you do not want network checks, avoid invoking the IP verification paths or remove/modify the SSH check; and (4) run it in a restricted account or sandbox if you want to limit any accidental disclosure. Overall: coherent and expected behavior for a verification checklist, but exercise the usual caution with tools that read local files and perform network checks.

Type: OpenClaw Skill Name: pre-task-checklist Version: 1.0.1 The skill bundle provides a structured verification protocol designed to improve agent reliability by preventing actions based on stale or hallucinated data. It includes a shell script (checklist.sh) and a Python utility (verify.py) that allow the agent to search local workspace files (e.g., TOOLS.md, AGENTS.md), verify file paths, and test SSH connectivity to known hosts. All code and instructions in SKILL.md and README.md are transparently aligned with the stated purpose of pre-task validation and lack any indicators of data exfiltration, persistence, or malicious prompt injection.