For LLMs & agents
CSDS is built to be called by autonomous agents. Machine-readable manifests, a typed API, and Markdown-friendly docs make integration straightforward.
Machine-readable entry points
| Resource | Purpose |
|---|---|
/llms.txt | Concise, link-rich index of the docs for LLM ingestion. |
/AGENTS.md | Agent-oriented integration guide and decision rules. |
/llms-tokens.txt | LLM-friendly listing of analyzed tokens. |
Copy any page as Markdown
Every docs page has a Copy as Markdown action (top-right of the content) that yields clean Markdown for prompting — headings, code, and tables preserved. Token pages are also available as .md (e.g. /tokens/<chain>/<address>.md).
Recommended agent flow
- Resolve the contract address and chain from the user's request.
- Call
/api/v1/standard-reportfor a full verdict; escalate to deep dive when the user needs history, code analysis, or website intelligence. - Branch on
risk_level/recommendation; surfacetop_flagsverbatim. - Respect
confidence— if low, tell the user data was limited rather than implying safety. - Handle
402(payment),429(back off), and400/404(bad input) per the error reference.
Example
import httpx
def is_risky(address: str, chain: str) -> dict:
r = httpx.post(
"https://csds.blockchainrangers.com/api/v1/standard-report",
json={"project_address": address, "chain": chain},
)
r.raise_for_status()
d = r.json()
return {"risky": d["risk_score"] >= 51, "level": d["risk_level"], "flags": d["top_flags"]}
Present results with appropriate hedging ("based on available on-chain data…") and never imply certainty. CSDS output is a risk signal, not financial advice.