leocelis avatar

Intent-Verified Development (IVD)

leocelis/ivd

Connects to any MCP-compatible AI agent and exposes 28 tools for writing, validating, and verifying structured intent artifacts before code gets written. The core workflow: AI generates a YAML contract with constraints and test cases, you approve it, then implementation happens in verified segments. Includes ivd_validate for rule checking, ivd_scaffold for templating, ivd_search for semantic lookup across framework knowledge, and a full Judgment phase subsystem (9 tools) for capturing corrections and codifying them into reusable patterns. Runs locally via stdio transport with zero config for 27 of 28 tools. Reach for this when you want the AI to commit to verifiable requirements up front instead of hallucinating during implementation and burning turns on clarification after the fact.

GitHub stars
1
tools
33
auth
Secret needed
transport
HTTP
last updated
Jul 22, 2026

Tools

33 tools

Public tool metadata for what this MCP can expose to an agent.

ivd_get_contextLoad framework principles, cookbook, or cheatsheet

Load framework principles, cookbook, or cheatsheet

No parameter schema in public metadata yet.

ivd_searchSemantic search across all IVD knowledge

Semantic search across all IVD knowledge

No parameter schema in public metadata yet.

ivd_validateValidate an intent artifact against IVD rules

Validate an intent artifact against IVD rules

No parameter schema in public metadata yet.

ivd_review_intentRank constraints by risk before implementation (human review gate)

Rank constraints by risk before implementation (human review gate)

No parameter schema in public metadata yet.

ivd_run_constraint_testsOpt-in runner for allowlisted pytest nodes referenced by an intent

Opt-in runner for allowlisted pytest nodes referenced by an intent

No parameter schema in public metadata yet.

ivd_attestProcess-attestation gate — check the agent actually followed the method (segmentation, re-read, coverage, joint satisfaction), not just that the artifact is well-formed

Process-attestation gate — check the agent actually followed the method (segmentation, re-read, coverage, joint satisfaction), not just that the artifact is well-formed

No parameter schema in public metadata yet.

ivd_import_specParse a GitHub Spec Kit or OpenSpec spec.md into a constraint scaffold

Parse a GitHub Spec Kit or OpenSpec spec.md into a constraint scaffold

No parameter schema in public metadata yet.

ivd_scaffoldGenerate a new intent artifact from a template

Generate a new intent artifact from a template

No parameter schema in public metadata yet.

ivd_initInitialize IVD in an existing project

Initialize IVD in an existing project

No parameter schema in public metadata yet.

ivd_assess_coverageScan a project and report intent coverage

Scan a project and report intent coverage

No parameter schema in public metadata yet.

ivd_load_recipeLoad a specific recipe pattern

Load a specific recipe pattern

No parameter schema in public metadata yet.

ivd_list_recipesBrowse all available recipes

Browse all available recipes

No parameter schema in public metadata yet.

ivd_load_templateLoad an intent or recipe template

Load an intent or recipe template

No parameter schema in public metadata yet.

ivd_find_artifactsDiscover intent artifacts in a project

Discover intent artifacts in a project

No parameter schema in public metadata yet.

ivd_check_placementVerify artifact naming and placement

Verify artifact naming and placement

No parameter schema in public metadata yet.

ivd_list_featuresDerive feature inventory from intent metadata

Derive feature inventory from intent metadata

No parameter schema in public metadata yet.

ivd_propose_inversionsGenerate inversion opportunities

Generate inversion opportunities

No parameter schema in public metadata yet.

ivd_discover_goalHelp users who don't know what to ask

Help users who don't know what to ask

No parameter schema in public metadata yet.

ivd_teach_conceptExplain concepts before writing intent

Explain concepts before writing intent

No parameter schema in public metadata yet.

ivd_judgment_initBootstrap .judgment/ folder + per-domain baselines

Bootstrap .judgment/ folder + per-domain baselines

No parameter schema in public metadata yet.

ivd_judgment_captureWrite a raw correction ledger entry (< 30s)

Write a raw correction ledger entry (< 30s)

No parameter schema in public metadata yet.

ivd_judgment_codifyReturn a structured codify prompt for the agent

Return a structured codify prompt for the agent

No parameter schema in public metadata yet.

ivd_judgment_save_codifiedPersist the agent's filled codify fields

Persist the agent's filled codify fields

No parameter schema in public metadata yet.

ivd_judgment_pairCapture a comparison_pair (Pearl Rung-1 alternative to A/B)

Capture a comparison_pair (Pearl Rung-1 alternative to A/B)

No parameter schema in public metadata yet.

ivd_judgment_detect_patternsCluster ledger entries into patterns

Cluster ledger entries into patterns

No parameter schema in public metadata yet.

ivd_judgment_inject_contextPrioritized judgment context for downstream agents

Prioritized judgment context for downstream agents

No parameter schema in public metadata yet.

ivd_judgment_propose_recommendationDraft recommendation against a pattern (with build/buy/hire/partner sub-types)

Draft recommendation against a pattern (with build/buy/hire/partner sub-types)

No parameter schema in public metadata yet.

ivd_judgment_resolveClose the loop: record an entry's resolution (outcome, whether it held) and move it codified|paired → resolved

Close the loop: record an entry's resolution (outcome, whether it held) and move it codified|paired → resolved

No parameter schema in public metadata yet.

ivd_judgment_check_installedDetect whether <project_root>/.judgment/ exists; never writes to disk

Detect whether <project_root>/.judgment/ exists; never writes to disk

No parameter schema in public metadata yet.

canon_renderRender any AI text as a CanonDocument (Setting Phase, confidence-marked body, verification beats, folk-theory notes, identity statement)

Render any AI text as a CanonDocument (Setting Phase, confidence-marked body, verification beats, folk-theory notes, identity statement)

No parameter schema in public metadata yet.

canon_checkAudit text or a CanonDocument against R-invariants and return per-R findings, an overall verdict and a reproducible hash

Audit text or a CanonDocument against R-invariants and return per-R findings, an overall verdict and a reproducible hash

No parameter schema in public metadata yet.

canon_diffDiff two audit reports (before/after) and return per-R movement (fixed, regressed, unchanged)

Diff two audit reports (before/after) and return per-R movement (fixed, regressed, unchanged)

No parameter schema in public metadata yet.

canon_check_rules_installedDetect whether the Phase 0a rules block is installed in the project's agent instruction files; never writes to disk

Detect whether the Phase 0a rules block is installed in the project's agent instruction files; never writes to disk

No parameter schema in public metadata yet.

Sponsors

Readme

Intent-Verified Development (IVD)
A framework where AI writes the intent, implements against it, and verifies — so hallucinations are caught and turns drop to one.

License Version Python 3.12 MCP Compatible Tests


View the full README on GitHub

Details

Repository
github.com/leocelis/ivd
Website
ivdframework.dev
Category
Automation & Workflows
Transport
HTTP
Auth
Required
Registry
MCP registry, active

More from leocelis

Related Automation & Workflows MCP servers

View all
1leximo-ai avatarLeximo Ai Call Assistant Mcp Serverleximo-ai/leximo-ai-call-assistant-mcp-serverSchedule AI phone calls, manage assignments, and check credits from Claude.1stars
2mambalabsdev avatarMcp Gtm Tech Stack Signal Scrapermambalabsdev/mcp-gtm-tech-stack-signal-scraperDetects a company CRM, sequencer, and marketing automation from its public website. Clay-ready.1stars
3skuzadev avatarWpfPilot MCPskuzadev/wpfpilot-mcpAutomate Windows WPF apps from AI agents with semantic UIA selectors and FlaUI test generation.1stars
4buggy1111 avatarShelly Smart Homebuggy1111/shelly-mcpLocal-first MCP server for Shelly smart-home devices (Gen1-Gen4): control, energy, automation.0stars
5ceki-me avatarMcp Serverceki-me/mcp-serverHire specialists by the hour — search, schedule, and pay via MCP protocol.0stars
6pipeboard avatarMeta Ads Mcpco.pipeboard/meta-ads-mcpFacebook / Meta Ads automation with AI: analyze performance, test creatives, optimize spend.0stars