procureai-skills
MCP server by AIprocurement.club · mcp
Fixed procurement workflows run a published sequence of stages; outputs become unverified context for later agents. Read structuredContent and isError, then inspect run.status or execution.status: a successful tool response can describe a running or failed saved run. Use procurement_runget to poll source-free saved status, waiting its pollAfterSeconds when running, and procurement_artifactget only when generated deliverable content is needed. Workflow summaries contain stage run IDs; retrieve each successful stage artifact separately. For a chain that may outlast your tool-call timeout, set runInBackground and poll procurement_workflow_executionget; approvalGates pause before named paid stages until procurement_workflow_executionapprove, which needs the buyer to agree first. Direct prepaid stages need their own supplied invoices. Enabled native marketplace billing instead requires verified platform authorization and a caller spending ceiling. Paid agent/workflow calls require…
Endpoint: https://mcp.aiprocurement.club/mcp
Verification: observed. Source: crawl.
Capabilities (47)
- xclause-risk-scoring
- xcontracts-summariser
- xnegotiate-coach
- xrenew-radar
- xmatch-supplier-shortlist
- xrfi-drafter
- xrfp-drafter
- xscore-bid-evaluation
- xcash-working-capital
- xmaverick-detector
- xsavings-narrative
- xspend-taxonomy
- xesg-scoring
- xmarket-briefs
- xrisk-monitor
- xvendorbench
- rfp-evidence
- finance-business-case
- negotiation
- knowledge
- mro
- indirect
- it-software
- it-hardware
- telecom