# Cory Parry > First-party project guides by Cory Parry, developer of Treebar and Intents. This file is a plain-text companion to the public project pages. It describes software and workflows, not published model benchmark results. ## Treebar - Page: https://coryparry.dev/projects/treebar - Updated: 2026-09-20 - Monitor Codex Git worktrees on macOS with Treebar. Explore the TreeCleaner preview for developer caches and merged worktrees, with setup and cleanup guidance. - Requirements: macOS 14 or later. Local Codex activity and Git worktrees to monitor, with access granted to the relevant folders. A physical notch is optional: the menu-bar view provides an alternative. ### What is a Git worktree, and why does Codex use one? A Git worktree is a separate working directory attached to the same repository. Each worktree can have its own checked-out branch, so you can work on different tasks without repeatedly switching the files in one checkout. Codex can use worktrees to keep parallel tasks separate. Treebar shows their recent activity and Git changes together on your Mac. Source: https://coryparry.dev/projects/treebar#what-is-a-worktree ### How can I monitor several Codex worktrees on a Mac? Install Treebar, grant access to your Codex data and repository folders, then open its notch or menu-bar view. It groups recent worktrees by repository and shows branch names, activity, additions and deletions. Open a checkout for changed files and more detail. Treebar follows local Codex activity; it is not a dashboard for every cloud agent or remote machine. Source: https://coryparry.dev/projects/treebar#monitor-codex-worktrees ### Why do Git worktrees take up space on my Mac? Linked worktrees share the repository’s Git data, but each has its own working files. Dependencies and build output created inside separate worktrees can also occupy additional space. Removing an unused worktree can recover its working-directory space; it does not necessarily remove shared caches elsewhere on the Mac. Source: https://coryparry.dev/projects/treebar#worktree-disk-space ### How does TreeCleaner help with Mac developer storage? The Treebar 1.0.4 preview includes TreeCleaner for reviewing developer caches and merged, inactive linked worktrees. Its categories include Xcode build data and caches, Swift package cache, Homebrew downloads, npm and pnpm caches. Review the scan before approving cleanup and granting temporary write access. This preview is not part of the 1.0.3 App Store release checked on 20 September 2026. It is a developer cleanup workflow, not a whole-disk duplicate finder. Source: https://coryparry.dev/projects/treebar#mac-developer-storage ### What should I check before removing an old worktree? Check that the task has stopped, that no process is using the directory, and that any work you need is committed and preserved. Review uncommitted and untracked files before removal. TreeCleaner’s preview workflow targets merged, inactive linked worktrees and preserves branches. Removing a worktree and deleting a branch are different operations; inspect both rather than assuming one implies the other. Source: https://coryparry.dev/projects/treebar#remove-old-worktrees ### Does Treebar read my prompts or change my repositories? Treebar monitors local activity metadata and Git status; it does not retain or display Codex prompts, messages or transcripts. Repository inspection is read-only. In the 1.0.4 preview, TreeCleaner is an explicit exception: cleanup needs your review and temporary write permission. Folder access remains under your control. Source: https://coryparry.dev/projects/treebar#treebar-privacy - [Treebar: current App Store release and compatibility](https://apps.apple.com/gb/app/treebar/id6794450399) - [Treebar product website](https://treebar.net) - [Git documentation: worktrees and removal](https://git-scm.com/docs/git-worktree) ## Intents - Page: https://coryparry.dev/projects/intents - Updated: 2026-09-30 - Evaluate Apple Foundation Models on Mac with repeatable test suites, scoring, traces and baseline comparisons. Connect coding agents through the built-in MCP server. - Requirements: macOS 27 or later and a Mac that supports Apple Intelligence. For the default on-device provider, enable Apple Intelligence and download its model. Xcode 27 is required to build from source or use Intent Lab; it is not required for normal on-device evaluations. ### How can I evaluate Apple’s Foundation Models on my Mac? Use Intents to create a suite of representative prompts, shared instructions and reference answers. Choose a scoring mode, run repeated trials, and inspect the responses and execution traces. Save the run as a baseline before changing your prompt or model settings. The default provider uses Apple’s on-device Foundation Model and requires a supported Mac with Apple Intelligence enabled and its model downloaded. Source: https://coryparry.dev/projects/intents#evaluate-apple-models ### Which scoring method should I use for an Apple AI evaluation? Use exact-text scoring when the entire expected response matters, contains-text scoring for required phrases, and an AI rubric for explicit quality criteria. Collect-only mode records responses and traces without assigning a score. Intents’ AI rubric uses a 1–4 scale, with 3 or 4 passing. Read the explanations and check reference answers: a rubric score is evidence for those test cases, not a general measure of model intelligence. Source: https://coryparry.dev/projects/intents#scoring-apple-models ### How do I compare model results and spot regressions? Run the same suite before and after a change, save both runs, and choose the earlier run as the baseline in Results and Compare. Inspect case-level responses, scores, timing and tool activity rather than relying only on an overall average. Keep the cases, scoring criteria and settings comparable, repeat variable tasks, and record your Mac and macOS version when sharing results. Source: https://coryparry.dev/projects/intents#compare-model-results ### Can Codex or another MCP-compatible agent run evaluations? Yes. Intents includes an MCP server for managing suites and references, running evaluations, inspecting traces and comparing saved results. For Codex, open Settings, choose Connect to Codex, restart Codex and keep Intents open. Other compatible clients can use the documented local HTTP endpoint at http://127.0.0.1:17873/mcp with the managed bearer credential. Intents stores this credential in the login Keychain and configures Codex to send it. The connector listens on the same Mac; authenticated local clients can read and change evaluation data. Source: https://coryparry.dev/projects/intents#agent-mcp-evaluations ### Where can I view or share Intents results? Review individual responses and traces in the app’s Results, compare saved runs, or export a JSON report. This page describes the evaluation tool; it does not publish a model leaderboard or claim benchmark scores. Before sharing an export, review it for prompts, responses, attachments and tool data. A useful report includes the suite, scoring criteria, model/provider, settings, repetitions and test environment. Source: https://coryparry.dev/projects/intents#evaluation-results ### Do Apple model evaluations run entirely on-device? The default on-device provider runs inference locally. Configured HTTP providers and tools receive the data needed for their calls, and optional Private Cloud Compute uses Apple’s network service when available. Saved runs and exported traces can contain sensitive content. Review the project’s data-handling guide before evaluating private documents or sharing reports. Source: https://coryparry.dev/projects/intents#local-inference - [Intents: installation, scoring and agent setup](https://github.com/coryparrry/Intents#readme) - [Download current Intents releases](https://github.com/coryparrry/Intents/releases) - [Apple: Foundation Models framework](https://developer.apple.com/documentation/foundationmodels)