19 docs tagged with "mcp"

Multi-tool sequences for complex problems — security plan, per-release checklist, change set analysis, findings sweep.

How to use the SbD-ToE MCP server to support the end-to-end agentic development process — the agent consults the manual via MCP at each stop of the flow, from mandate to kill-switch.

Organisational policy that defines the treatment of AI models, datasets, MCP servers/tools, embedded prompts and AI providers as an auditable supply chain — AI BOM generation per build in a standardised format (CycloneDX 1.6 ml-bom), version pinning, a list of approved providers and response to upstream incidents, complementing Policy 10 (dependencies) and Policy 11 (SBOM), proportionate to the criticality level (L1, L2, L3).

How to label answers generated with the MCP — manual-grounded vs observed vs inferred vs not verified — and the most common mistakes to avoid.

Applying plan_sbd_toe_repo_governance to a new repository — list of artefacts by chapter and automatic scaffold.

Generating code with prepare_sbd_toe_codegen_context — branching by status, citations discipline, mandatory security_rationale.

Detailed installation instructions for the SbD-ToE MCP in every supported MCP client.

Official MCP server of the SbD-ToE — connecting Claude, Copilot, Cursor and other MCP clients to the Manual with deterministic grounding.

Cross-referencing SbD-ToE with the AI Act, CRA, DORA, NIS2, GDPR — with the MCP (canon and indexed cross-checks) and the web manual for content later than the snapshot.

get_guide_by_role to generate personalised onboarding guides — practices by SDLC phase + user stories joined.

Automated review of Pull Requests against the active controls of the risk_level — findings citable by CTRL-*.

Connecting the SbD-ToE MCP to Claude Code or Cursor in under a minute.

MCP resources (sbd://toe/*) and pre-packaged prompts of the SbD-ToE MCP — when to use each one.

Configuring the AI client (Claude, Copilot, Cursor) to consult the SbD-ToE automatically — via skill files, agent files, persistent instructions.

Combining get_threat_landscape + consult_security_requirements by concern to produce grounded threat models.

Every tool of the SbD-ToE MCP — parameters, deterministic vs heuristic semantics, input/output examples, and when to prefer one over another.

Symptoms, causes and solutions for common problems with the SbD-ToE MCP — content lag, versions, debugging, integration with AI clients.

Ready-made recipes combining tools, resources and prompts of the SbD-ToE MCP — PR audit, grounded codegen, threat modelling, governance bootstrap, onboarding and normative cross-check.

Versioning policy of the SbD-ToE MCP, relationship with Manual versions, and public roadmap.