Agent Memory Kit

PassAudited by ClawScan on May 1, 2026.

Overview

This appears to be a purpose-aligned agent memory framework, but users should notice that it creates persistent reusable memory and includes optional shell-based search setup.

Install this if you want a persistent local memory system for an agent. Before using it, decide what belongs in the memory folder, avoid storing secrets, review memory files periodically, and treat the optional search CLI/PATH setup as executable code rather than pure markdown documentation.

Findings (3)

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.

Low

#ASI06: Memory and Context Poisoning

What this means

Anything saved in the memory files may be reused by the agent in future sessions.

Why it was flagged

The skill intentionally makes persistent local memory part of the agent's recurring context, which is central to the kit but can cause stale, sensitive, or incorrect notes to influence later work.

Skill content

Add to your `AGENTS.md` ... Before doing anything else: 1. Read `SOUL.md` ... 3. Read `memory/YYYY-MM-DD.md` ... 4. **If in MAIN SESSION**: Also read `MEMORY.md`

Recommendation

Keep the memory folder scoped, review it periodically, and avoid putting passwords, private tokens, or untrusted instructions into persistent memory.

Low

#ASI05: Unexpected Code Execution

What this means

Using the optional search feature can execute local scripts from the skill package.

Why it was flagged

The documentation tells users to add a skill-provided command to their shell PATH. This is purpose-aligned for the search feature, but it means local shell code may be run by name.

Skill content

export PATH="$PATH:$HOME/.openclaw/workspace/skills/agent-memory-kit/bin" ... memory-search --help

Recommendation

Only add the command to PATH if you intend to use the search feature, and review the script source before relying on it.

Low

#ASI04: Agentic Supply Chain Vulnerabilities

What this means

A user following the fallback instructions could update or add code outside the reviewed package contents.

Why it was flagged

The install guide includes a user-directed unpinned update from a remote branch, and the provided manifest does not list the referenced `bin/memory-search` wrapper.

Skill content

If missing, pull latest Memory Kit:

```bash
cd skills/agent-memory-kit
git pull origin main
```

Recommendation

Prefer the packaged registry version, or pin and review any GitHub source before running updated scripts.