ClawHub

Agent Memory Architecture

v1.0.0

Complete zero-dependency memory system for AI agents — file-based architecture, daily notes, long-term curation, context management, heartbeat integration, a...

⭐ 2· 347·1 current·1 all-time
by@1kalin
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
VirusTotalVirusTotal
Benign
View report →
OpenClawOpenClaw
Benign
high confidence
✓
Purpose & Capability
Name and description promise a zero-dependency, file-based memory system and the SKILL.md only describes managing markdown files (MEMORY.md, ACTIVE-CONTEXT.md, memory/*.md, heartbeat-state.json). There are no unrelated environment variables, binaries, or external APIs required, so the requested capabilities align with the stated purpose.
✓
Instruction Scope
Runtime instructions are focused on local file organization, read/write rules, retention and curation policies, progressive loading, and a simple heartbeat file. The instructions do not direct network exfiltration, unknown endpoints, or access to unrelated system config. They do require filesystem read/write access within the agent workspace (auto-create memory/ directory, append daily notes), which is coherent for a file-based memory system.
ℹ
Install Mechanism
No install spec or code files — lowest-risk, instruction-only skill. Minor documentation mismatch: README shows a 'clawhub install afrexai-agent-memory-system' command and slash-commands (/memory-status, /remember) that imply an installable package or runtime helpers, but the registry entry contains only SKILL.md/README.md and no installable package. This is a documentation/marketing inconsistency rather than an active risk, but users should not expect additional binaries to be installed by this skill.
✓
Credentials
The skill declares no required environment variables or credentials. SKILL.md explicitly recommends not storing secrets in memory files and recommends limiting MEMORY.md exposure to main sessions; the scope of requested access (workspace files) is proportionate to its stated function.
✓
Persistence & Privilege
always:false and user-invocable:true (defaults) — no elevated persistence requested. The skill does not ask to modify other skills or system-wide settings. Its behaviors are limited to reading/writing files in the agent workspace as described.
Assessment
This skill is an instruction-only guide for a file-based memory system and is internally coherent. Before installing/using it: (1) confirm your agent runtime grants filesystem write/read only to a safe, isolated workspace — these instructions require creating and editing local files; (2) do not put secrets, API keys, or passwords into the memory files (the SKILL.md explicitly warns against this); (3) be aware README references an install command and slash-commands that are not present in the package — treat those as marketing/optional extras, not built-in functionality; (4) if you want operational helpers (commands, automation), ask the author for an actual package/source code and review it before running; (5) if privacy is a concern, consider encrypting sensitive notes or restricting where the workspace is stored/backed up.

Like a lobster shell, security has layers — review code before you run it.

agentvk9778y4amqynvrdgs7k3wxpa3x821an5contextvk9778y4amqynvrdgs7k3wxpa3x821an5knowledgevk9778y4amqynvrdgs7k3wxpa3x821an5latestvk9778y4amqynvrdgs7k3wxpa3x821an5memoryvk9778y4amqynvrdgs7k3wxpa3x821an5openclawvk9778y4amqynvrdgs7k3wxpa3x821an5ragvk9778y4amqynvrdgs7k3wxpa3x821an5

License

MIT-0
Free to use, modify, and redistribute. No attribution required.
Termshttps://spdx.org/licenses/MIT-0.html

Comments