Tiered Memory

v1.0.1

Two-tier memory system for OpenClaw agents. Tier 0 = QMD semantic search for recent memories (7-14 days). Tier 1 = SQLite archive for long-term storage. Auto...

0· 57·0 current·0 all-time
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/description, required binaries (python3, ollama), and the included scripts all align: scripts scan ~/.openclaw/workspace/memory, summarize with Ollama, insert into a local SQLite DB, and move files to an archive directory. The requested binaries and filesystem paths are appropriate for a memory-archiver skill.
Instruction Scope
SKILL.md and the CLI/python examples limit actions to reading/writing files under ~/.openclaw, indexing/searching those files, and optionally invoking Ollama for summarization. The cron instructions explicitly run the provided archiver script. There are no instructions to read unrelated system config, other users' data, or to post content to unknown external endpoints.
Install Mechanism
This is instruction-only (no automated installer). The README suggests installing Ollama using the project's official install script (curl | sh) and pulling a model; that's a documented, optional step and uses the vendor's URL. No arbitrary archive downloads or obscure URLs are used by the skill itself. As always, running third-party install scripts has inherent risk and should be reviewed before execution.
Credentials
The skill requests no credentials or environment variables and confines work to ~/.openclaw paths and a local SQLite DB — proportionate to its purpose. One consideration: the archiver passes file content to the 'ollama' CLI. If Ollama is configured to run local models, content remains local; if an Ollama installation is configured to proxy or send data externally, sensitive memory content could be transmitted. The skill offers --skip-llm to avoid LLM summarization.
Persistence & Privilege
always is false and the skill does not request persistent platform-wide privileges. It creates and uses files under the user's home (~/.openclaw) only, and the cron job in SKILL.md is an explicit user action (not automatic). It does not attempt to modify other skills or system-wide agent settings.
Assessment
This skill appears to do what it claims: auto-archive old OpenClaw memory files into a local SQLite DB and optionally summarize them with Ollama. Before installing, check: 1) you are comfortable that the archiver will read and move files under ~/.openclaw/workspace/memory and will create ~/.openclaw/memory_archive.db; 2) if you care about privacy, use the --skip-llm flag or ensure your Ollama installation runs entirely locally (so summaries are not proxied off-machine); 3) inspect and, if desired, run the scripts manually before adding the cron entry; and 4) avoid running third-party curl|sh installers without review (the README suggests one for Ollama). If those points are acceptable, the skill is coherent and appropriate for its stated purpose.

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

latestvk979gqecg71p303899gdjsmhz183wdnf

License

MIT-0
Free to use, modify, and redistribute. No attribution required.

Runtime requirements

Binsollama, python3

Comments