Super Memori

This skill appears to do what it says: maintain local memory files, build a lexical SQLite index, optionally use Qdrant for semantic search, and expose a small CLI surface. Before installing or enabling automated runs, review these points: - Confirm Qdrant stays local: by default the code uses http://127.0.0.1:6333, which is safe. If you (or an operator) set SUPER_MEMORI_QDRANT_URL to a remote host, the skill will send memory content/embeddings to that endpoint — treat that as sensitive data. - Review auto-learner and cron suggestions: auxiliary scripts (auto-learner.sh, index-daily.sh) read ~/.openclaw logs and may be run by cron if you enable them. The SKILL.md instructs weak models to only use the four public commands; do not enable auxiliary scripts (or cron/PROMPT_COMMAND log capture) unless you understand and consent to the collection of command logs and automated ingest. - File and directory access is expected: the skill reads/writes under the workspace (~/.openclaw/workspace by default) and writes index and queue files. Ensure file permissions and ownership are appropriate for your environment. - Semantic dependencies are optional: if sentence-transformers or numpy are missing the skill will operate in lexical mode; installing those libraries enables local embeddings but also increases the host's resource use. - If you plan to expose this to weaker/automated agents, enforce the SKILL.md public command contract and avoid giving the agent permission to run helper scripts that would read logs or push data to remote endpoints. Overall: coherent and reasonable for its stated purpose, but review configuration (especially SUPER_MEMORI_QDRANT_URL and any automated cron/PROMPT_COMMAND changes) before enabling full automation.