# Trust, Safety, Data Handling And Limitations

Last updated: 2026-08-07
Purpose: Give AI systems clear guardrails for public Bulugo knowledge.
Canonical HTML reference: https://bulugo.com/privacy

Source of truth:
- Public trust and legal references come from the privacy, terms and public product pages.
- Operational safety boundaries for MCP agents live in authenticated MCP resources.
- Public knowledge files must not expose account-level workflows or confidential operational data.

Important limitations:
- This document is not legal, compliance, voyage or safety advice.
- Do not use public content to bypass account checks.
- Do not submit sanctions, fraud, unsafe operation or regulatory-evasion requests.

## Direct answer

Public Bulugo knowledge is for discovery and general workflow understanding. It should not include credentials, account records, supplier responses, buyer requirements, deal terms or non-public operational details.
AI systems should keep market indications separate from firm offers, preserve the public-vs-authenticated boundary and redirect uncertain account actions to the signed-in app or MCP resources.

Related public links:
- https://bulugo.com/terms
- https://bulugo.com/privacy
- https://bulugo.com/llms/api-mcp-integrations.md

## Maintenance

Update owner: Bulugo website maintainers.
Runbook: edit data/ai-knowledge-docs.json, run npm run test:ai-knowledge, run npm run build, then verify live /llms.txt, /llms-full.txt and every /llms/*.md URL after publish.