← Back to Skills Marketplace
80
Downloads
0
Stars
0
Active Installs
1
Versions
Install in OpenClaw
/install super-session-cleanup
Description
Session cleanup skill for Claw-family agents (OpenClaw, WorkBuddy, QClaw, etc.). This skill should be used when the user wants to track and clean up temporar...
Usage Guidance
This package appears to do what it says: it creates a workspace .workbuddy/session-track.json, records temporary files and installs, and can delete files and uninstall packages. Before installing or enabling:
- Review scripts/cleanup.py (especially send_to_trash and uninstall paths) and run in --dry-run mode first to see what would be removed.
- Understand that when tracking is enabled the skill will silently inspect tool calls and append entries to the session-track.json file — it does not transmit data externally but will record paths and package names.
- Be cautious about system-wide uninstall steps (pip uninstall, apt/winget/choco/brew) since those require elevated privileges and can remove shared dependencies; the skill marks such actions as requiring confirmation in SKILL.md, but confirm prompts depend on how your agent is configured.
- If you want to avoid any autonomous changes, disable autonomous invocation for this skill or only invoke it manually and use the script with --dry-run first.
Overall: coherent and consistent with its described purpose; test in a safe workspace first and inspect the track file before executing cleanup.
Capability Analysis
Type: OpenClaw Skill
Name: super-session-cleanup
Version: 1.4.0
The 'super-session-cleanup' skill is a legitimate utility designed to track and remove temporary artifacts (files, packages, and skills) created during an agent session. The core logic in `scripts/cleanup.py` includes robust safety guardrails, such as a `FORBIDDEN_ROOTS` list to protect sensitive directories (Home, Desktop, Documents, etc.) and a 'trash-first' deletion strategy that utilizes OS-native recycle bins before attempting hard deletes. The bundle is exceptionally well-documented and includes extensive functional, regression, and lifecycle test suites, indicating a high level of engineering quality and no evidence of malicious intent or data exfiltration.
Capability Assessment
Purpose & Capability
Name/description match the included artifacts and code. The repo provides a CLI (scripts/cleanup.py), helper docs, and SKILL.md that describe tracking and removing temp files, pip/npm packages, system packages, and skills — which aligns with a 'session-cleanup' purpose.
Instruction Scope
SKILL.md asks the agent to record artifacts to a persistent <workspace>/.workbuddy/session-track.json and to perform a mandatory post-check after every tool call while tracking is active (scanning write_to_file/execute_command strings and silently appending). That behavior is consistent with tracking but is broad: it gives the skill automatic, background access to examine tool calls and record file paths without announcing each append. It does not instruct network exfiltration or reading unrelated secret env vars, but the silent, automatic nature is a privacy/behavioral concern users should be aware of.
Install Mechanism
No install spec or remote downloads; the package is instruction-only from the registry perspective and ships local Python scripts and docs. No external installers or extracted archives are fetched at install time.
Credentials
The skill requests no environment variables or credentials. The code uses standard OS utilities (PowerShell/osascript/gio/trash-put, pip, npm) and filesystem paths, which are proportionate to a cleanup tool. There are no requests for unrelated tokens or secrets.
Persistence & Privilege
always:false (normal). The skill creates a persistent tracking file in the workspace and expects autonomous post-checking after tool calls while tracking is enabled. Autonomous invocation is allowed by default (disable-model-invocation:false); that combined with mandatory, silent background updates means the skill can autonomously record session artifacts. This is expected for a cleanup skill but users should be aware that tracking occurs silently while enabled.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install super-session-cleanup - After installation, invoke the skill by name or use
/super-session-cleanup - Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.4.0
super-session-cleanup 1.4.0 changelog:
- Added automatic tracking of all session-generated artifacts (temp files, scripts, packages, skills, etc.), with persistent storage in session-track.json.
- Introduced two-phase lifecycle: Phase 1 (tracking, with triggers) and Phase 2 (cleanup, with triggers).
- Cleanup procedure now displays a categorized, Markdown-formatted checklist of tracked items, and excludes any explicitly retained items.
- Safe-to-delete items are removed automatically; items with dependencies prompt for confirmation before uninstalling or deletion.
- New auto-tracking rules ensure thorough capture of all relevant resources after each tool call.
- Provides batch or individual deletion strategies, with OS-specific safe delete commands for items outside the workspace.
Metadata
Frequently Asked Questions
What is session-cleanup?
Session cleanup skill for Claw-family agents (OpenClaw, WorkBuddy, QClaw, etc.). This skill should be used when the user wants to track and clean up temporar... It is an AI Agent Skill for Claude Code / OpenClaw, with 80 downloads so far.
How do I install session-cleanup?
Run "/install super-session-cleanup" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is session-cleanup free?
Yes, session-cleanup is completely free, licensed under MIT-0. You can download, install and use it at no cost.
Which platforms does session-cleanup support?
session-cleanup is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created session-cleanup?
It is built and maintained by Chaobs (@chaobs); the current version is v1.4.0.
More Skills