Manage quark-auto-save(QAS, 夸克自动转存, 夸克转存, 夸克订阅) tasks via CLI.
Security checks across static analysis, malware telemetry, and agentic risk
Overview
This appears to be a real QAS management helper, but it can make broad changes to cloud files/tasks and persist configuration-derived habits, so users should review its scope before use.
Install only if you trust the configured QAS server and are comfortable letting the agent manage QAS tasks and cloud-file operations. Before use, require confirmation for deletes, renames, config changes, and run-all actions, and decide whether QAS task/path habits may be saved in TOOLS.md.
Static analysis
No static analysis findings were reported for this release.
VirusTotal
VirusTotal findings are pending for this skill version.
Risk analysis
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.
If invoked incorrectly, the agent could delete or rename cloud files, change QAS settings, or trigger many saved tasks at once.
The client exposes destructive and broad account-mutating actions, including cloud-file deletion/config updates and a no-argument run mode that runs all tasks, without artifact evidence of confirmation or scope guards.
python3 qas_client.py delete-file <path> # Delete cloud file ... python3 qas_client.py update-config <json> # Update config ... # Mode 1: Run all tasks (empty data)
Require explicit user confirmation for delete, rename, update-config, delete-task, and run-all actions; prefer dry-run previews and require specific task/path arguments.
The configured QAS endpoint receives a token that can authorize task and file operations; URL query tokens may also appear in server or proxy logs.
The QAS token is required and used to authenticate API requests, which is expected for this integration, but it is appended to request URLs.
QAS_TOKEN = os.environ.get("QAS_TOKEN", "") ... query = f"token={QAS_TOKEN}"Use only a trusted QAS_BASE_URL, prefer HTTPS where possible, protect logs, and use the least-privileged token available.
Private task names, media organization patterns, or save paths could be stored in persistent agent context and reused in later interactions.
The skill directs the agent to retrieve configuration/task information and persist derived saving habits in TOOLS.md, but does not define filtering, retention, or user review boundaries.
After the user sets the token, the following analysis must be performed and recorded in TOOLS.md: ... get-config ... Extract `savepath` directory patterns from existing tasks ... Record to TOOLS.md
Ask the user before writing QAS-derived information to TOOLS.md, store only minimal non-sensitive preferences, and provide a way to review or remove the stored notes.