PLS Agent Tools

Provides utilities for safe file handling, JSON/YAML editing, regex text processing, system commands, encoding, date/time, and validation tasks.

This skill is essentially a cookbook of shell commands that align with its stated purpose. Before installing or enabling it for autonomous use: (1) verify your agent runtime has the referenced tools (jq, yq, rsync, trash, lsof, curl, etc.) or adapt examples to your environment; (2) be aware examples include destructive operations (deleting files, killing processes) and network requests—run them in a sandbox or with backups and least-privilege accounts first; (3) if you don't want the agent to run these commands automatically, disable autonomous invocation for this skill or require explicit user invocation; (4) review the SKILL.md and remove or modify any commands that would operate on sensitive paths or send data to external endpoints you don't control.

Type: OpenClaw Skill Name: pls-agent-tools Version: 1.0.0 The skill bundle provides extensive shell command execution capabilities, including arbitrary file system manipulation (read, write, delete), process management (`kill -9`), and critical network operations via `curl`. Specifically, `curl -O URL` allows downloading arbitrary files from external sources, and `curl -X POST ... URL` allows sending arbitrary data to external endpoints. While presented as general utilities, these capabilities create a significant attack surface for prompt injection against the AI agent, potentially enabling remote code execution, data exfiltration, or system compromise if the agent is instructed maliciously. There is no explicit evidence of intentional malice (e.g., hardcoded malicious URLs or specific exfiltration targets) within the `SKILL.md` itself, but the broad and powerful nature of the exposed commands makes it highly risky.