Qiuqiu Helper

Automates workspace tasks including summarizing recent changes, adding timestamped notes, and cleaning old log files.

This skill could be useful, but it currently leaves important details unspecified. Before installing or enabling it, ask the author to: (1) provide a clear description of purpose and intended scope; (2) explicitly declare/configure the exact directories it will write to (memory folder) and clean (logs directory); (3) implement and document safety checks (limit file patterns, require user confirmation before deletion, provide a dry-run mode, avoid recursive deletes by default); (4) restrict operations to a workspace-specific directory and run with least privilege; and (5) add audit/logging so changes are visible. If the author cannot clarify these points, treat the skill as risky because it could delete or overwrite files unexpectedly.

Type: OpenClaw Skill Name: qiuqiu-helper Version: 0.1.0 The `SKILL.md` defines tools with high-risk capabilities that are susceptible to prompt injection. Specifically, `quick_note` allows writing to a 'specified file' and `clean_logs` allows deleting files from a 'specified directory'. While the descriptions imply benign defaults and restricted scopes ('in the memory folder', 'current logs directory'), the parameters (`file`, `path`) are broad. A malicious prompt could instruct the AI agent to use path traversal (e.g., `../../../../etc/passwd`) or target critical system directories (e.g., `/`) for file write or deletion, leading to data corruption, privilege escalation, or system compromise. The skill itself does not contain explicit malicious instructions, but its design exposes significant potential for abuse by a compromised agent.