通过 WeKnora REST API 导入文档,并对知识库执行向量+关键词混合检索。
记忆
Claw SQLite Knowledge
试用Knowledge base skill that uses the published clawsqlite CLI for ingest, search, show, and maintenance workflows.
它能做什么
Knowledge base skill that uses the published clawsqlite CLI for ingest, search, show, and maintenance workflows.
技能文档
ClawSQLite Knowledge Skill
This skill is a thin wrapper around the published clawsqlite PyPI package.
Important: installing this skill from ClawHub installs only the wrapper files.
The clawsqlite CLI is not usable until bootstrap_deps.sh has been run.
After bootstrap, prefer the stable skill-local entry:
/skills/clawsqlite-knowledge/bin/clawsqlite
It does not:
- vendor the
clawsqlitesource tree - clone any Git repository
- define a second runtime layer on top of the official CLI
- redefine a parallel JSON API
It does:
- install the pinned
clawsqlite==1.0.12PyPI package throughbootstrap_deps.sh - guide agents to use the official
clawsqlite knowledge ...CLI - document common workflows for knowledge-base operations
Knowledge Instance Home
This skill directory is only the instruction shell. Do not store the user's
knowledge DB or private config here. Run clawsqlite knowledge ... from a
knowledge instance home instead. In OpenClaw, use the persistent workspace data
directory:
mkdir -p ~/.openclaw/workspace/data/clawsqlite-knowledge/default
cd ~/.openclaw/workspace/data/clawsqlite-knowledge/default
On a normal non-OpenClaw Linux install, the default falls back to
${XDG_DATA_HOME:-~/.local/share}/clawsqlite-knowledge/default.
The local private config must be:
./clawsqlite.toml
clawsqlite.toml is the single runtime configuration source. Do not create a
second config file, do not rely on shell environment variables for normal
configuration, and do not guess DB paths.
Bootstrap
Check or install the required published package:
sh bootstrap_deps.sh
Managed Python environments may not expose a global clawsqlite command on
PATH. Use bin/clawsqlite from this skill directory; it wraps the installed
PyPI package and local fallback target.
Then create or edit ./clawsqlite.toml inside the knowledge instance home. If
no config exists yet:
bin/clawsqlite knowledge maintenance init-config --instance default
cd ~/.openclaw/workspace/data/clawsqlite-knowledge/default
Validate
After installation and config editing, validate with:
/skills/clawsqlite-knowledge/bin/clawsqlite knowledge maintenance doctor --json
Doctor is lightweight by default. Only pass --check-llm or
--check-embedding when the user explicitly wants provider roundtrip checks.
Agent Rules
- Use only the official
clawsqliteCLI. - Stay in the knowledge instance home when running
clawsqlite knowledge .... - Use the three-level command tree:
record,maintenance,analysis. - Do not call removed flat commands such as
clawsqlite knowledge ingest. - Do not vendor, clone, or patch
clawsqliteinside this skill directory. - Do not use degraded ingest unless the user explicitly asks for it.
- Report
ERROR,ERROR_KIND, andNEXTlines from the CLI directly.
Common Commands
Strict URL ingest:
/skills/clawsqlite-knowledge/bin/clawsqlite knowledge record ingest \
--url "https://example.com/article" \
--category web_article \
--json
Strict text ingest:
/skills/clawsqlite-knowledge/bin/clawsqlite knowledge record ingest \
--text "A thought worth saving." \
--title "A saved thought" \
--category thought \
--json
--title is a source_title hint for archive filenames and metadata. In
strict ingest, the stored knowledge title is generated_title from the
configured LLM.
Search:
/skills/clawsqlite-knowledge/bin/clawsqlite knowledge record search "vector database design" --mode hybrid --topk 5 --json
Show one record:
/skills/clawsqlite-knowledge/bin/clawsqlite knowledge record show --id 123 --full --json
Maintenance:
/skills/clawsqlite-knowledge/bin/clawsqlite knowledge maintenance reindex --check --json
/skills/clawsqlite-knowledge/bin/clawsqlite knowledge maintenance cleanup --days 3 --dry-run --json
/skills/clawsqlite-knowledge/bin/clawsqlite knowledge maintenance backup --dry-run --json
Analysis:
/skills/clawsqlite-knowledge/bin/clawsqlite knowledge analysis build-interest-clusters --json
/skills/clawsqlite-knowledge/bin/clawsqlite knowledge analysis report-interest --days 7 --no-pdf
Explicit degraded ingest, only when the user requested it:
/skills/clawsqlite-knowledge/bin/clawsqlite knowledge record ingest \
--text "Local no-network test." \
--title "Test" \
--gen-provider off \
--allow-heuristic \
--allow-missing-embedding \
--json
相关技能
Install and configure the security-related plugins required by OpenClaw, including the `claw-sentry` plugins. Use this skill when you want to complete instal...
Audit an OpenClaw agent workspace and generate standardized evaluation reports, scores, and patches. Use when asked to review memory quality, retrieval effic...
Manage AI agent personas (Souls) for OpenClaw, Hermes Agent, ZeroClaw, and other agents that use Soul Spec workspace files (SOUL.md, IDENTITY.md, AGENTS.md). Installs/switches/restores persona files in the agent workspace, searches and publishes to the clawsouls.ai registry (network + token auth), optionally syncs encrypted agent memory via Git (explicit `sync`/`swarm` commands only), and detects the local agent platform. Use only when the user explicitly asks to manage personas — e.g. "install a soul", "switch persona to X", "list souls", "restore my previous soul", "publish my soul", or "login to clawsouls". Workspace-modifying and publishing actions should be confirmed with the user first.
Use when an AI Agent (Claude Code, Codex, OpenClaw, or similar) needs to operate an llm-wiki knowledge base: ingest source files into Markdown wiki pages, answer questions from wiki/index.md and linked pages, run agent-bridge status/lint/link/relink/merge/query/index tasks, preserve provenance and t
集成公告源监控、签名校验与风险技能处置的 ClawSec 套装管理入口。