← Back to Skills Marketplace
raaipro

Knowledge Base Pro

by RAAIPRO · GitHub ↗ · v3.5.3 · MIT-0
cross-platform ⚠ suspicious
95
Downloads
0
Stars
0
Active Installs
2
Versions
Install in OpenClaw
/install raai-knowledge-base-pro
Description
Corporate knowledge brain for Russian-speaking teams. Replaces HR onboarding department + internal communications team on 40-60% of routine. FAQ bot, role-ba...
Usage Guidance
This package appears to be a coherent corporate KB product, but there are multiple mismatches you should verify before installing or running anything: 1) Inspect install.sh and test/smoke-test.sh before executing them; run them only in an isolated VM or container and never as root. 2) The SKILL.md metadata claims "no network requests" while the product explicitly supports OpenAI/Anthropic embeddings and Notion/Confluence/Google Drive & Telegram integrations — assume network activity if you populate config.yaml with those connectors. 3) Only provide API keys (OPENAI_API_KEY, ANTHROPIC_API_KEY, NOTION_API_KEY, CONFLUENCE_API_TOKEN, GOOGLE_SERVICE_ACCOUNT_JSON, TELEGRAM_BOT_TOKEN) if you trust the code and have audited the install/test scripts; otherwise use local storage mode and avoid embedding models until you review network calls. 4) If you need deeper assurance, request the content of install.sh and test/smoke-test.sh (or run a static review) and verify there are no unexpected remote downloads, credential uploads, or commands that read unrelated file paths. 5) Consider running the smoke-test in a network‑restricted environment first to observe behavior safely. These inconsistencies make the skill 'suspicious' rather than clearly benign; a quick code review of the install/test scripts would likely resolve the ambiguity.
Capability Analysis
Type: OpenClaw Skill Name: raai-knowledge-base-pro Version: 3.5.3 The bundle is a highly detailed and well-documented corporate knowledge management skill for the OpenClaw platform, specifically localized for Russian-speaking teams. It provides comprehensive instructions for an AI agent to manage FAQs, onboarding tracks, and Standard Operating Procedures (SOPs). The included shell scripts (install.sh, build.sh, and smoke-test.sh) perform standard file management and validation tasks without any evidence of malicious behavior, such as unauthorized network calls, credential theft, or persistence mechanisms. The requested environment variables (e.g., NOTION_API_KEY, OPENAI_API_KEY) are appropriate for the tool's stated purpose of integrating with external knowledge sources and providing semantic search capabilities.
Capability Tags
cryptocan-make-purchasesrequires-sensitive-credentials
Capability Assessment
Purpose & Capability
The name/description (corporate KB, local/no‑SaaS operation, Telegram bot optional) align with the files (config.yaml with local/notion/confluence/google_drive options, onboarding, indexing, embeddings). However there are conflicting claims: SKILL.md repeatedly states "works without external SaaS" while the config and examples are built to support Notion/Confluence/Google Drive, embedding_model set to text-embedding-3-large (OpenAI), and Telegram bot integration. Those integrations are plausible for a KB product, but the marketing/metadata overstates "no external SaaS" while the implementation expects optional external APIs.
Instruction Scope
SKILL.md instructs filling config.yaml and running install.sh / test/smoke-test.sh per README. The skill's operation (indexing → chunk → embedding, analytics, exports, Telegram bot) inherently requires network calls to external APIs when configured (OpenAI/Anthropic/Notion/Confluence/Google Drive/Telegram). Yet the SKILL.md metadata lists network_behavior.makes_requests: false and uses_agent_telegram: false — an explicit mismatch with the described runtime actions. The instructions also imply reading config.yaml and using API keys if provided; that is expected but the mismatch in stated network behavior is a red flag. The runtime instructions do not appear to instruct arbitrary system-wide reads beyond the project files, but they do ask the operator to run included shell scripts (install.sh, smoke-test), which can run arbitrary commands.
Install Mechanism
There is no install spec in the registry (instruction-only), but the package contains build.sh, install.sh and test/smoke-test.sh and README recommends running install.sh and smoke-test.sh. That means local scripts are used for installation by the operator — a normal delivery mechanism for a packaged product, but also higher risk because those scripts may run arbitrary shell commands. No otherwise-suspicious remote download URLs were found in the visible build.sh, and build.sh itself is a benign packaging script. The lack of an explicit, reviewed install spec combined with recommended execution of shipped install scripts raises a safety concern: inspect install.sh/test scripts before running, and run in an isolated environment.
Credentials
The registry lists no required env vars; SKILL.md enumerates optional credentials (ANTHROPIC_API_KEY, OPENAI_API_KEY, NOTION_API_KEY, CONFLUENCE_API_TOKEN, GOOGLE_SERVICE_ACCOUNT_JSON, TELEGRAM_BOT_TOKEN). Those variables are proportionate to the product's optional integrations (embeddings/models, Notion/Confluence/Google Drive ingestion, Telegram bot). The concern is the default embedding_model (text-embedding-3-large) and many example flows assume use of cloud APIs — the skill may offer a local mode but full functionality seems to require cloud keys. The metadata claiming no network requests contradicts the need for those keys when integrations are used.
Persistence & Privilege
The skill does not request 'always: true' and does not declare any system-wide config path requirements. There is no evidence it modifies other skills or global agent settings. Autonomous invocation is allowed (platform default) but is not combined with other high‑privilege flags. This dimension shows no unusual privilege escalation.
How to Use
  1. Make sure OpenClaw is installed (local or Docker)
  2. Run the install command in chat: /install raai-knowledge-base-pro
  3. After installation, invoke the skill by name or use /raai-knowledge-base-pro
  4. Provide required inputs per the skill's parameter spec and get structured output
Version History
v3.5.3
Marketplace presentation refresh.
v3.5.1
Initial marketplace release.
Metadata
Slug raai-knowledge-base-pro
Version 3.5.3
License MIT-0
All-time Installs 0
Active Installs 0
Total Versions 2
Frequently Asked Questions

What is Knowledge Base Pro?

Corporate knowledge brain for Russian-speaking teams. Replaces HR onboarding department + internal communications team on 40-60% of routine. FAQ bot, role-ba... It is an AI Agent Skill for Claude Code / OpenClaw, with 95 downloads so far.

How do I install Knowledge Base Pro?

Run "/install raai-knowledge-base-pro" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.

Is Knowledge Base Pro free?

Yes, Knowledge Base Pro is completely free, licensed under MIT-0. You can download, install and use it at no cost.

Which platforms does Knowledge Base Pro support?

Knowledge Base Pro is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).

Who created Knowledge Base Pro?

It is built and maintained by RAAIPRO (@raaipro); the current version is v3.5.3.

💬 Comments