Claude Code Memory Kit
v1.0.0Stop Claude Code from repeating mistakes — enforce guardrails, preserve context, maintain consistency across sessions
⭐ 0· 12·0 current·0 all-time
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Benign
high confidencePurpose & Capability
The name/description (prevent repeat mistakes, enforce guardrails) matches the actual contents: templates (CLAUDE.md, memory.md, mistakes.md, bootstrap.md), installation guidance, and guardrail configuration suggestions. There are no unexpected binaries, credentials, or network dependencies declared.
Instruction Scope
SKILL.md, INSTALL.md, and templates only instruct the agent/user to copy template files, configure ~/.claude/settings.json, run the 'claude' CLI, and keep small memory files. That is within scope. Two items to be aware of: (1) bootstrap.md mentions 'Database: [connection details if needed]' and 'API endpoints: [base URLs]' — these are placeholders but could encourage storing sensitive connection strings in project or bootstrap files if users are not careful; (2) TROUBLESHOOTING.md suggests settings like {"permissions":{"allow":["Read"]}} and gives an npm global install example with --unsafe-perm=true — these are benign guidance in context but should be handled cautiously. Overall the instructions do not read or exfiltrate data and do not reference unexpected external endpoints.
Install Mechanism
No install spec and no code files that will be executed. This is instruction-only (templates and docs) so nothing will be downloaded or written by an installer. Low install risk.
Credentials
The skill declares no required env vars, no credentials, and no config paths beyond recommending user-managed locations (project root files and optional ~/.claude/settings.json). The requested scope is proportional to the stated functionality.
Persistence & Privilege
The skill is not always-enabled and does not request elevated platform privileges. It suggests local configuration of Claude Code (per-user config and project files) which is appropriate for persistent behavior. Nothing indicates modification of other skills or system-wide settings beyond the user-editable Claude config.
Assessment
This is a documentation-and-template package, not executable code — that lowers risk. Before installing or committing the templates: 1) Review the templates (CLAUDE.md, memory.md, mistakes.md, bootstrap.md) for any hardcoded secrets, API keys, or DB credentials and never commit secrets to the repository. 2) When following INSTALL.md and TROUBLESHOOTING.md, avoid blindly running global installs from unknown packages and avoid using --unsafe-perm unless you understand the implications. 3) The kit suggests adding preToolGates and permission entries to ~/.claude/settings.json — check what those settings do in your Claude Code client and confirm you’re comfortable granting the listed permissions. 4) Test the guardrails in a safe environment (non-production repo) to ensure they behave as expected. Overall the package appears consistent with its stated purpose, but practice standard review and secret-management hygiene before deploying team-wide.Like a lobster shell, security has layers — review code before you run it.
latestvk971jmp4r8vxnhbrm2wn0ha95984a7cx
License
MIT-0
Free to use, modify, and redistribute. No attribution required.