Memoria Persistente
v1.0.0Sistema de memória hierárquica em 3 camadas para agentes IA: curto prazo (contexto), médio prazo (diário), longo prazo (conhecimento). Tudo em português.
⭐ 0· 114·0 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
Name and description describe a 3-layer persistent memory system and the SKILL.md only requires creating/reading files under a .memory/ directory — that is coherent. Minor inconsistency: the README shows CLI usage examples (memoria-persistente salvar/recuperar) and an install line (clawhub install memoria-persistente), but the package is instruction-only and includes no binary or install spec; consumers should not assume a separate executable will be installed.
Instruction Scope
Instructions are narrowly scoped to creating and reading local files in .memory/ (context, daily, knowledge). They do not reference external endpoints, unrelated system files, or secret env vars. However, the instructions are vague about handling sensitive data, encryption, access controls, retention, or file permissions — a privacy risk if the agent persists secrets or PII.
Install Mechanism
There is no install spec and no code files; this is the lowest-risk pattern. The SKILL.md includes an example 'clawhub install' command but the registry metadata contains no install script or binary — likely informational rather than an actual download/install step.
Credentials
The skill declares no required environment variables, no credentials, and no config paths. That is proportionate for a local file-based memory system.
Persistence & Privilege
Flags are standard (always: false, user-invocable: true, model invocation allowed). The skill does not request elevated privileges or permanent/forced inclusion.
Assessment
This skill is instruction-only and appears to implement a local .memory/ folder layout for short/medium/long-term memories. Before using it: (1) Understand that any persisted memories will be written to disk in the agent's working directory (.memory/) — avoid storing secrets or PII there. (2) Ensure proper file permissions or encryption if you will persist sensitive info. (3) The SKILL.md shows CLI commands and a 'clawhub install' example, but no executable or install spec is included; don't expect a separate binary unless the publisher provides one. (4) Test the skill in an isolated environment to confirm behavior and where files are written. If you need stronger privacy guarantees, request guidance from the author on encryption/retention or implement an explicit policy to avoid persisting sensitive data.Like a lobster shell, security has layers — review code before you run it.
latestvk97czgnmsba6070p864xzpj20x83bbmd
License
MIT-0
Free to use, modify, and redistribute. No attribution required.