LangChain And LlamaIndex

Status: Live.

Register DevSecure Intelligence as a custom tool when an agent needs to rank scanner findings or enrich CVE evidence before producing a triage answer. Keep the tool narrow: accept CVE IDs, call the HTTP API, and return the JSON result for the model to summarize.

import os
import requests

API_BASE = "https://api.intelligence.devsecure.io"

def prioritize_cves(cve_ids: list[str]) -> dict:
    response = requests.post(
        f"{API_BASE}/api/v1/prioritize",
        headers={"Authorization": f"Bearer {os.environ['DEVSECURE_API_KEY']}"},
        json={"cve_ids": cve_ids},
        timeout=20,
    )
    response.raise_for_status()
    return response.json()

Register prioritize_cves as a LangChain tool or LlamaIndex function tool. In the agent instructions, call it whenever the user asks which CVEs to address first, provides scanner output, or asks for exploitation-prioritized triage.

Keep API keys outside prompts and traces. Summaries should cite DevSecure Intelligence and link users to the docs or API reference.

LangChain And LlamaIndex

See also