ClawHub

Claude Code Memory Skill

v0.1.0

Use when you want to set up, maintain, or review a Claude Code style layered memory workflow, including `CLAUDE.md` rules, session memory, durable memory, an...

⭐ 0· 20·0 current·0 all-time
byLiu Wenyu@indulgeback
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
Capability signals
CryptoCan make purchases
These labels describe what authority the skill may exercise. They are separate from suspicious or malicious moderation verdicts.
VirusTotalVirusTotal
Pending
View report →
OpenClawOpenClaw
Benign
high confidence
✓
Purpose & Capability
The name and description (Claude Code style layered memory) match the content of SKILL.md and accompanying docs. The skill is instruction-only and only needs access to repository files (instruction file and .agent-memory), which is appropriate for its stated purpose.
✓
Instruction Scope
Runtime instructions direct the agent to read and review local files (CLAUDE.md, .agent-memory/MEMORY.md, session summary) and produce promotion/cleanup reports. The SKILL.md explicitly instructs not to modify durable/instruction files without user approval. It does not ask the agent to read unrelated system paths, environment variables, or send data to external endpoints.
✓
Install Mechanism
No install spec or code files are executed — this is instruction-only, which is the lowest-risk install mechanism and consistent with the skill's purpose.
✓
Credentials
The skill requires no environment variables, credentials, or external configuration paths. All declared and implicit requirements are proportional to a repo-local memory review workflow.
✓
Persistence & Privilege
The skill is not forced-always and does not request elevated or persistent platform privileges. It can be invoked by the agent (default), which is expected for skills; the SKILL.md also restricts automatic modification of memory/instruction files without explicit user approval.
Assessment
This skill is a documentation/workflow template for organizing repo-local memory and appears coherent. Before installing or enabling it: (1) inspect any .agent-memory files in your repo for secrets or sensitive data (durable memory is stored in repo files and could contain tokens if someone put them there), (2) confirm you want the agent to read those repo files (it will read CLAUDE.md/.agent-memory to produce reports), and (3) avoid giving the agent autonomous write permission to instruction or durable memory unless you trust it and have review controls in place. Because it is instruction-only and declares no external endpoints or credentials, the main risk is accidental storage/exposure of sensitive info in the .agent-memory files — review those files before use.

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

ai-agentsvk97asavn5d2sdegyhpk2jxh0q984dn16claude-codevk97asavn5d2sdegyhpk2jxh0q984dn16developer-toolsvk97asavn5d2sdegyhpk2jxh0q984dn16latestvk97asavn5d2sdegyhpk2jxh0q984dn16memoryvk97asavn5d2sdegyhpk2jxh0q984dn16

License

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

Comments