otito - Executive Summary¶
Purpose: Independent trust and merge evidence for AI-assisted software teams
Maintainer: Oluwasegun Olumbe
Version: See package.json in the repository root for the canonical version; server.json is kept in sync for the MCP manifest. The surface includes multi-domain discoverability, eval + data-access subcommands, code maps for C#, Python, Java, Ruby, and Rust, and the v1.0 absorption (impact, pass, pass-pr, review).
Overview¶
otito answers one practical question:
It is a Node.js CLI and MCP server that produces deterministic, local-first repository context and merge evidence for coding agents, maintainers, and reviewers. Native agent harnesses generate the change. Otito proves whether it is safe to merge.
What We Have Defined¶
| Area | Summary |
|---|---|
| Repository inspection | Files, package metadata, languages, package managers, scripts, entrypoints, and git state |
| Code maps | JSON-first source maps with multi-domain tagging, imports, exports, symbols, routes, and language extractors for TS/JS, Go, C#, Python, Java, Ruby, Rust |
| Context packs | Task-aware file suggestions, related files, patterns, tests, validation commands, vendor-bundle filtering, and data-access boosts |
| Data-access surface | Detects inline SQL and Prisma ORM calls, aggregated by source / operation / table / file |
| Eval suite | Measures token savings, retrieval/risk accuracy, inferred command execution, and real local-gate effectiveness against committed fixtures |
| Local catalog | Discovery, indexing, and search across local repositories |
| PR review context | Diff-aware review prompts, changed domains, risk flags, Go test-file detection, and optional GitHub comments |
| MCP support | Agent-callable tools for repo inspection, maps, search, harnesses, workspaces, and PR review |
| Governance | CI gates, Òtítọ́ readiness, CODEOWNERS, SemVer guidance, security reporting, templates, and review policy |
| Demo packet | Company-facing packet that links the executive summary, case study, proof run, launch note, and pilot checklist |
| Pilot runbook | Step-by-step first repository and pull request pilot with roles, evidence, stop conditions, and triage |
| Thesis pack | Public explainer pages for harness, tutorials, convergence, usage, determinism, dual-mode, prompt determinism, the trust harness, and clean code |
| Proof index | Sanitized public evidence map plus private/internal evidence boundaries for company reviewers |
| Review policy | Branch protection, required checks, CODEOWNERS, conversation-resolution, and admin-decision snapshot |
| Feedback loop | Structured public pilot intake that turns reviewer concerns into docs, gates, proof, or roadmap work |
| Company adoption | Screenshot-style case study for evaluation, pilot rollout, and governance decision records |
| Public launch note | Short external-facing story for Òtítọ́, proof runs, and next trust-layer gates |
| Operating loop | Repeatable session rhythm for context, focused change, visible gates, human decisions, and durable evidence |
Product Position¶
otito is the independent trust layer beside native agent harnesses:
Models generate the change. Otito proves whether it is safe to merge.
native hosts -> generation (Codex, Claude Code, Gemini, Cursor)
otito -> context before change, evidence before merge
Humans -> accountability before release
This makes otito useful for maintainers who want AI-assisted development without losing the shape of the repository, test expectations, ownership boundaries, and review discipline. It does not compete with generic orchestration. Deterministic merge evidence is the differentiated core; organisation-level governance is the commercial edge.
Gate Signals¶
| Signal | Current Handling |
|---|---|
| Unknown repository shape | otito repo surfaces package, language, script, and git state |
| Unclear task scope | otito context generates task-aware primary and related files |
| Agent invents extra structure | Context packs and CONTRIBUTING keep the change in the smallest owner files |
| Multi-repo context missing | otito workspace builds a product-level context report |
| PR review surface hidden | otito pr summarizes changed files, risks, prompts, and comments |
| Merge readiness hidden | Òtítọ́ PR readiness runs on pull requests and records owner-decision warnings |
| Context evidence disconnected | Òtítọ́ reports include Context Evidence commands for context packs and PR review reports |
| Agent tool integration needed | otito mcp exposes repository context through MCP |
| MCP setup unclear | MCP workflow docs include generic stdio, Claude Desktop, VS Code, and Cursor examples |
| Contributor readiness | CONTRIBUTING.md, Code of Conduct, CI, CODEOWNERS, templates, security, release docs, and branch protection guidance |
| Review policy visibility | Review-policy snapshot summarizes branch protection and owner/admin decision boundaries |
| Gate behaviour needs proof | otito eval --gate-effectiveness checks one valid control and six expected blocks against the real local gate |
| Trust-layer demo | Public walkthrough for Òtítọ́ context, PR review context, and human merge accountability |
| Session continuity | Builder-founder operating loop keeps proof, gates, and next actions outside chat memory |
Critical Items Requiring Decision¶
| Priority | Item | Action Required |
|---|---|---|
| Medium | Package distribution | Decide npm registry publication path or continue GitHub install first |
| Medium | Demo assets | Keep dated proof runs, launch notes, and company adoption evidence current |
| Medium | Company pilot | Run the workflow with a real external reviewer and record feedback |
Next Steps¶
- Keep the company adoption packet current as Òtítọ́ evolves.
- Run one real repository and pull request through the company pilot runbook.
- Keep the proof index current as public artifacts and private proof boundaries evolve.
- Capture feedback from real company reviewers through the pilot feedback loop and turn it into docs, gates, proof, or roadmap work.
- Use the builder-founder operating loop for long-running sessions so context, gates, decisions, and next actions remain inspectable.
- Keep MCP client examples current as Claude Desktop, VS Code, Cursor, and other hosts evolve.
- Extend the committed gate corpus when a real pilot reveals a new false pass, false block, or unclear reason.