dsh-web
zhu1090093659/dsh-webDeepSeek Harness(DSH)Web 插件聚合生态包 · 一切皆插件,创意工坊分发(dsh-market.com)
Chemical safety intelligence for AI-assisted experiment design.
An MCP server that gives AI agents (Claude, ChatGPT, Claude Code, Gemini CLI, Copilot Studio, and any other MCP client) 24 tools for chemical safety reasoning: compatibility checks, hazard analysis, regulatory compliance, PPE, storage, transport, and signed audit reports. Every answer names the supplier SDS it came from, with a link to the original PDF.
MSDS Chain, powered by ChainSDS: a verified, always-current chemical safety database.
Built for researchers who design experiments with AI and want safety verification inside the workflow rather than as a separate step: whether the chemicals on a deck are compatible, which mixing orders are dangerous, what PPE a specific handling step needs, how a compound is treated under EU REACH or US OSHA/TSCA, and what a GLP/GMP reviewer will ask for afterwards.
Hosted endpoint. Nothing to install, nothing to run locally:
claude mcp add msds-chain --transport http https://mcp.lagentbot.com/mcpThen run /mcp in Claude Code and sign in. A browser opens, you enter your email, and you paste back the code we send you. Your account is provisioned automatically and the free plan includes a credit allowance. There is no API key to create beforehand.
On another platform? See Platform Setup below.
Tool · Description `batch_safety_check` · One-call report for a chemical list: pairwise compatibility + per-chemical risk warnings `check_regulatory_lists` · Cross-reference a chemical against 23 regulatory watch lists across 8 jurisdictions + 3 international conventions `get_sds_section` · Retrieve a specific SDS section (1-16) for a chemical `get_sds_document` · Signed download URL (~5 min) for the original SDS/MSDS PDF; includes source provenance `get_chemical_alternatives` · Safer substitutes for restricted or high-risk chemicals `validate_protocol_chemicals` · Extract & validate chemical names from protocol text or code `check_mixing_order` · Safe addition sequence for reagent pairs (e.g., acid into water) `get_waste_disposal` · Waste classification, container type, and disposal procedures `upload_msds_pdf` · Upload MSDS PDF for AI-powered parsing and data extraction (requires API key) `draft_sds_sections` · Draft Sections 8 and 15 of an SDS for your own mixture from its ingredient list; every line cites its source, and missing ingredients are listed rather than guessed `compare_sds_versions` · Hazard-change diff between two SDS versions (H-code additions/removals + whether they change a verdict) check_chemical_compatibility · Pairwise compatibility for 2+ chemicals get_chemical_risk_warnings · GHS classification, H-codes, signal words, flash point get_ppe_recommendation · Gloves, eye protection, respiratory, body protection get_storage_guidance · Storage class, cabinet type, temperature, isolation rules get_emergency_response · Spill, fire, or exposure emergency procedures get_exposure_limits · OEL/TLV/PEL/MAC across US, EU, JP, CN, INT get_transport_classification · UN number, hazard class, packing group, ADR/IATA/IMDG check_regulatory_compliance · Multi-region compliance status, list-backed for EU, US, CN, JP, CA, AU, SG search_chemical_database · Look up chemicals by name, synonym, or CAS number ask_chemical_safety · Natural language catch-all for any safety question create_audit_session · Full audit with signed PDF report (requires API key) get_audit_report · Download link for the signed audit PDF
Experiment protocol review:
User: I'm planning a Grignard reaction with magnesium turnings, diethyl ether,
and bromobenzene. Check if this setup is safe.
Agent:
→ calls batch_safety_check(["magnesium", "diethyl ether", "bromobenzene"])
→ Returns: compatibility matrix + per-chemical risk warnings (with source SDS)Signed audit report:
User: Create a safety audit report for our quarterly review.
Chemicals: acetone, methanol, ethanol, isopropanol, hexane.
Agent:
→ calls create_audit_session("Q2 2026 Solvent Cabinet Review", [...])
→ calls get_audit_report("SESSION-ID")
→ Returns: signed PDF URL (Ed25519 signature, suitable for GLP/GMP compliance)One hosted server, two transports. Pick the row your client supports:
Transport · Endpoint · Notes Streamable HTTP · https://mcp.lagentbot.com/mcp · Preferred: broadest and most robust SSE · https://mcp.lagentbot.com/sse · For clients that only speak SSE
Authentication is either OAuth (the client opens a browser sign-in on first connect) or a static header, Authorization: Bearer sk-msds-your-key, for headless clients. Create a key at msdschain.lagentbot.com > API Keys.
Settings > Connectors, search msds-chain, and enable it. Sign-in happens on first use.
Three routes, any one is enough:
# 1. Remote server only
claude mcp add msds-chain --transport http https://mcp.lagentbot.com/mcp
# 2. Plugin from the official Claude plugin marketplace (server + skill)
/plugin install msds-chain@claude-plugins-official
# 3. npm shim (registers the same hosted endpoint)
claude mcp add msds-chain -- npx -y msds-chain-mcp@latestManual config in ~/.claude.json:
{
"mcpServers": {
"msds-chain": {
"type": "http",
"url": "https://mcp.lagentbot.com/mcp"
}
}
}Routes 2 and 3 of the plugin also install the /msds-safety-check skill: it spots chemicals in your conversation and offers a check, and runs a guided audit for lab protocols or EHS compliance.
Listed in the OpenAI Plugin Directory. Search MSDS Chain in ChatGPT, enable it, and sign in when prompted. Works on web, desktop, and mobile.
Listed in the Gemini CLI extensions gallery:
gemini extensions install https://github.com/littleblakew/msds-chain-mcpGemini asks you to confirm twice: once to trust the workspace, once to review the extension and the skill it carries. Your first chemical question starts the sign-in automatically; the server supports Dynamic Client Registration, so Gemini registers itself and completes the OAuth exchange on its own. Check the connection with gemini mcp list, then just ask in chat:
Can acetone and sodium hypochlorite be stored in the same cabinet?codex marketplace add https://github.com/littleblakew/msds-chain-mcpMSDS Chain ships a certified Power Platform connector, so there is nothing to configure by hand. In Copilot Studio, add a tool from the connector directory and search MSDS Chain; the connector handles the OAuth exchange against the same hosted endpoint.
These clients do not speak streamable HTTP yet, so point them at /sse and pass a static key:
https://mcp.lagentbot.com/sseAuthorization : Bearer sk-msds-your-keyUse the endpoint table above. Clients that implement OAuth 2.1 with Dynamic Client Registration need no pre-registration; everything else uses the static Authorization header.
upload_msds_pdf, create_audit_session, get_audit_report) associate activitywith your account for audit traceability.
database per your account's data-sharing settings.
Industry-sourced, AI-verified, and cryptographically signed.
plus international conventions
you relied on
Symptom · Likely cause · Fix 401 Unauthorized on connect · Missing or invalid API key · Check the Authorization: Bearer sk-msds-... header. Generate a fresh key at msdschain.lagentbot.com > API Keys. Tools don't appear in the client · Server not loaded · Fully restart the client after adding the server, then confirm msds-chain shows in the MCP/tools list. Connection drops or times out · Wrong transport · Prefer https://mcp.lagentbot.com/mcp; SSE sessions can drop when the server redeploys. GET https://mcp.lagentbot.com/health should return {"status":"ok"}. Credit balance exhausted · Credit allowance used up · The error says how to continue: an upgrade link and, on the free plan, when credits reset. Empty results for a known chemical · Name not matched · Retry with the CAS number or a common synonym; search_chemical_database accepts name, synonym, or CAS. OAuth sign-in fails · OAuth metadata unreachable · Confirm GET https://mcp.lagentbot.com/.well-known/oauth-authorization-server returns 200; the client should auto-discover the authorize and token endpoints.
Still stuck? Email contact@lagentbot.com or open an issue at github.com/littleblakew/msds-chain-mcp/issues.
MIT
Built by LAgentBot, AI-powered chemical safety infrastructure.
Part of the MSDS Chain platform: the world's first AI Agent-driven chemical safety data trust network, powered by ChainSDS.