← Back to Skills Marketplace
jiuge897

Memory Pruner

by jiuge897 · GitHub ↗ · v1.0.0 · MIT-0
cross-platform ⚠ suspicious
271
Downloads
0
Stars
0
Active Installs
1
Versions
Install in OpenClaw
/install memory-pruner
Description
Manages agent memory by auto-pruning old or low-relevance entries, compressing duplicates, and reporting storage costs with safe defaults.
Usage Guidance
Do not install or run this skill without verifying where the 'memory-pruner' CLI comes from. The SKILL.md instructs file reads and deletions but the package does not include the referenced executable or memory/ directory; an agent would rely on a local or external binary of unknown provenance. Before using: (1) inspect and obtain the memory-pruner implementation from a trusted source, (2) run only dry-run/analyze first and inspect suggested deletions, (3) ensure backups are stored in a safe location you control, (4) restrict the agent’s permissions so it cannot delete unrelated files, and (5) prefer manual confirmation before any prune operation. If you cannot verify the CLI origin or prefer not to grant filesystem deletion rights, consider declining this skill.
Capability Analysis
Type: OpenClaw Skill Name: memory-pruner Version: 1.0.0 The skill bundle describes a utility for managing and optimizing agent memory by pruning old or low-relevance entries. The documentation (SKILL.md) and configuration (config.json) are consistent with this stated purpose and contain no evidence of malicious intent, data exfiltration, or prompt injection. While the primary executable script 'memory-pruner' was not included in the provided files, the metadata and instructions appear legitimate and safe.
Capability Assessment
Purpose & Capability
The name, description, and commands (prune, analyze, compress, dry-run) coherently describe a memory management tool. However, the SKILL.md and config.json reference a 'memory-pruner' CLI and a 'memory/' working directory as part of the skill's files, yet the package contains no executable or code files. That leaves the agent instructed to run a binary that is not provided by the skill bundle.
Instruction Scope
The runtime instructions explicitly instruct reading, analyzing, compressing, backing up (.bak), and deleting workspace memory files. Those operations are within the claimed purpose, but because this is an instruction-only skill with no implementation, an agent following these instructions could run any local 'memory-pruner' binary present on the system or an attacker-supplied binary. The SKILL.md promises 'never deletes without confirmation', but instructions still include destructive actions (prune) and file I/O — the absence of an included implementation and of strict path constraints raises risk.
Install Mechanism
No install spec is provided (lowest install risk). However, config.json lists 'entry':'memory-pruner' and 'runtime':'shell', implying an entrypoint that does not exist in the bundle. This mismatch means the skill expects an external CLI to be present; the origin and integrity of that CLI are undefined.
Credentials
The skill requests no environment variables, no credentials, and no config paths. That is proportionate to the stated purpose. There is no request for unrelated secrets or external service keys.
Persistence & Privilege
The skill is not always-enabled and uses normal autonomous invocation defaults. It does not request persistent system-wide configuration changes in its bundle. The main concern is not privilege flags but the potential for destructive filesystem operations if an external 'memory-pruner' binary is invoked.
How to Use
  1. Make sure OpenClaw is installed (local or Docker)
  2. Run the install command in chat: /install memory-pruner
  3. After installation, invoke the skill by name or use /memory-pruner
  4. Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.0
Initial release of Memory Pruner—intelligent memory management for agents. - Auto-prunes old memories based on configurable thresholds. - Scores and retains more relevant memories, merging similar ones for efficiency. - Tracks memory usage costs and provides actionable analysis and recommendations. - CLI includes prune, analyze, compress, config, and dry-run commands. - Features safe deletion (confirmation and backup) and category-based organization.
Metadata
Slug memory-pruner
Version 1.0.0
License MIT-0
All-time Installs 0
Active Installs 0
Total Versions 1
Frequently Asked Questions

What is Memory Pruner?

Manages agent memory by auto-pruning old or low-relevance entries, compressing duplicates, and reporting storage costs with safe defaults. It is an AI Agent Skill for Claude Code / OpenClaw, with 271 downloads so far.

How do I install Memory Pruner?

Run "/install memory-pruner" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.

Is Memory Pruner free?

Yes, Memory Pruner is completely free, licensed under MIT-0. You can download, install and use it at no cost.

Which platforms does Memory Pruner support?

Memory Pruner is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).

Who created Memory Pruner?

It is built and maintained by jiuge897 (@jiuge897); the current version is v1.0.0.

💬 Comments