v1.2.0 — Claims vs Reality Verification + MCP Server

What's New

Claims vs. Reality Verification

A verify_claims stage that extracts factual claims from agent output, matches each against the signed provenance record, and classifies them as verified, unverified, contradicted, or fabricated. The verification report is itself a signed artifact — meta-provenance.

MCP Server

SpineFrame is now an MCP server (10 tools, stdio + SSE). Any MCP client (Claude Desktop, Cursor, custom agents) can run audits, retrieve reports, and verify provenance. Published on official MCP registry as io.github.Ratnaditya-J/spineframe.

MCP-Based Evidence Gathering

Compliance pipeline gathers evidence from GitHub and filesystem via MCP tool-use loop with dynamic turn allocation.

Other Changes

• Cross-platform artifact hashing fix (Windows compatibility)
• Compliance report provenance and stats
• Cost propagation from tool-use loops to run manifest
• 62 registered stage types, 930 tests

v1.1.2 — MCP Server + Registry

What's New

MCP Server

SpineFrame is now an MCP server. Any MCP client (Claude Desktop, Cursor, custom agents) can invoke SpineFrame as a tool.

pip install spineframe[mcp]
spineframe-mcp --runs-dir ./runs --config providers.json

10 tools: run_compliance_audit, run_osint_recon, run_research, get_run_status, list_runs, get_report, get_artifacts, get_provenance, fork_run, verify_run

Published to Official MCP Registry

io.github.Ratnaditya-J/spineframe on registry.modelcontextprotocol.io

Other Changes

• MCP-based evidence gathering for compliance pipeline (GitHub + filesystem sources)
• Cross-platform artifact hash fix (Windows compatibility)
• PyPI package now includes full README