ClawTrust
4.5k Downloads
Overview
Provides a multi-layer, token-efficient interface for accessing ClawdBot documentation using local golden snippets, a JSON search index, cached full pages, and on-demand web_fetch retrieval with freshness (TTL) tracking.
Key Advantages
1.Highly token-efficient by prioritizing local golden snippets and index lookups before any remote fetch.
2.Clear 3-layer architecture: snippets → index → full-page fetch, with explicit examples for each step.
3.Leverages native web_fetch instead of external MCP/brightdata, reducing cost and latency and removing external dependencies.
4.Freshness-aware via TTL per page category, helping balance cache reuse vs. up-to-date docs.
5.Built-in synonym support in the search index to improve fuzzy keyword matching across languages/phrasing (e.g., configuration/config/nastavenie).
Use Cases
- Answering common ClawdBot operational questions from pre-cached golden snippets (e.g., Telegram setup, allowFrom issues, OAuth errors, update procedure).
- Guiding users through provider/channel setup and gateway configuration using indexed documentation pages.
- Troubleshooting ClawdBot (restart procedures, OAuth problems, configuration basics) with minimal token usage.
- Discovering and reading less-common or new documentation topics via search index + web_fetch fallback (e.g., multi-agent concepts, skills tooling).
- Performing low-cost, fast documentation lookups in production ClawdBot deployments where latency and token usage matter.
Evaluation Scores
8.4
/ 10
Reliability
7.5
Functionality
8.5
Usability
8.8
Safety
9.5
Performance
9.0
Compatibility
6.5
Based on 1 evaluation · Latest: 3/19/2026
Download Trend
Loading...
Evaluation History (1)
8.4/103/19/2026▼
OS: linux-x64LLM: google/gemini-3-flash-preview
**Verdict:** Strong, well-structured documentation access skill for ClawdBot environments, with a clear focus on token efficiency and latency. Excellent fit if you are actively running ClawdBot and want fast, cheap access to its docs; much less useful outside that ecosystem.
**What it does well**
- Implements a sensible 3-layer strategy (golden snippets → search index → fetch) that should cover 80–90% of queries from local data.
- Uses native `web_fetch` instead of external scraping MCPs, improving speed, cost, and operational simplicity.
- TTL-based freshness per page category and a JSON index with synonyms provide decent relevance and staleness control.
- Documentation is concrete and example-driven, including common Slovak/English query patterns and explicit file paths/commands.
**Key risks / limitations**
- **Ecosystem lock-in:** Hard dependency on ClawdBot-specific paths (`~/clawd/data/...`) and the `web_fetch` core tool; portability to other OpenClaw setups or different bots is low.
- **Partial coverage:** Only ~50 core pages indexed and 7 golden snippets; edge-case or newer topics may still require full-page fetches or manual discovery via `llms.txt`.
- **Version/metadata mismatch:** The fetched docs mention v2.x, while provided local metadata says 1.0.0; consumers should verify they’re using the latest deployed version and that changelog features (e.g., web_fetch migration) are actually available in their instance.
- **Reliance on external docs site:** If `https://docs.clawd.bot` is down or changes structure, full-page fetch and index refresh may break unless the skill includes robust error handling (not described here).
**Recommended scenarios**
- You operate ClawdBot and need **fast, low-cost, mostly-local** answers to routine setup, config, and troubleshooting questions.
- You want an **LLM-friendly docs workflow** that minimizes token usage, with a clear path for fallbacks when snippets/index miss.
- You’re optimizing a ClawdBot-based support or ops assistant that frequently references official docs and benefits from TTL-based caching.
Not recommended if you’re not in a ClawdBot context or if you need a general-purpose documentation searcher across arbitrary sites or tools.
Comments (0)
No comments yet. Be the first!