connorch

connorch/skills

Skill Claude CodeCodex

Use when the user asks to monitor, watch, or babysit a PR.

connorch/skills

Skill Codex

Implements scoped code changes with Codex CLI, then inspects the diff. This is how the GPT model rated intelligence 9 in the model-selection table is invoked for implementation work. Use when the user asks to delegate implementation to Codex or a GPT model, when the model-selection rubric routes the work to that…

connorch/skills

Skill Codex

Gets an independent code review from Codex CLI. This is how the GPT model rated intelligence 9 in the model-selection table is invoked for review work. Use when the user asks to have Codex or a GPT model review work, when the model-selection rubric calls for that model's review perspective, or when Codex should audit…

connorch/skills

Skill Claude CodeCodex

Use when the user asks to file, open, or create a PR.

connorch/skills

Skill Claude CodeCodex

Writes HTML writeups and publishes them privately to files.wovn.org. Use when the user asks for an HTML writeup of work, such as a plan, spec, findings, report, or UI mocks, that is not part of the codebase.

connorch/skills

Skill Claude CodeCodex

Sync a forked repository with its upstream. Use when asked to "pull upstream", "sync the fork", "merge upstream", "update the fork", or "catch up with upstream".

connorch/skills

Skill Codex

Writes an end-to-end QA plan from the branch diff without running tests. Use when asked to create an end-to-end QA or UX verification plan for the current Git branch, especially from branch diffs, changed files, product flows, routes, screens, components, API changes, copy, styles, or config.

connorch/skills

Skill Codex

Runs a QA plan in a browser and reports the results without fixing anything. Use when asked to execute, validate, test, or browser-check a QA guide, QA plan, PR QA checklist, UX test plan, browser-based product test plan, or output from the $qa-ux-plan skill.

connorch/skills

Skill Codex

Step back from a branch that grew through iteration - review rounds, point fixes, PR feedback - and judge the end state as a whole: are the big architecture calls right, are abstractions drawn in the right places, is logic shared when appropriate, and did the churn leave residue (dead code, redundant guards, stale…

connorch/skills

Skill Claude CodeCodex

Use whenever the user provides a link to files.wovn.org, when a local file needs a public or private URL, or when the user asks to upload, host, share, list, or retrieve files through Wovn.

connorch/skills

Skill Codex

Use when the user approves a plan and says to implement it. Takes the plan through implementation, browser QA, and the PR review loop to a PR that is ready to look at.

connorch/skills

Skill Codex

Delegate a bounded task to Claude Code CLI (claude -p) as a headless subagent, then review what it produced. Use when the user asks to hand work to Claude or Claude Code, when a multi-file implementation, refactor, investigation, or independent review would benefit from a second coding agent, or when heavy work should…

connorch/skills

Skill Claude CodeCodex

Use when the user asks to 3D print something, find or design a model to print, check a model for printability, slice it, check on the Bambu printer or the AMS, or see what is on the print bed. Covers finding models on MakerWorld and Printables, generating them with AI, making them from a script, the Printability…

connorch/skills

Skill Codex

Run commands on another of the user's machines (the macOS and linux machines on their tailnet) over Tailscale SSH. Use when the user asks to check, run, or change something on a different machine, such as "make sure main is pulled on my MacBook Pro" or "how much disk is free on the Studio".

connorch/skills

Skill Claude CodeCodex

Use when the user asks to use Jev or Kev, when a task needs bulk classification, yes/no filtering, or scoring of many items, or when browser work is repetitive or high-volume (many searches, collecting results across pages). Jev is TypeSafe's fast decision model and drives the jevx CLI and the Agents Chrome on this…

connorch/skills

Skill Claude CodeCodex

Use when a task needs a GUI app on a Mac whose desktop is locked, or when the user asks to start, stop, or set up the agent macOS VM. Boots an on-demand Tart VM with its own always-unlocked desktop, then shuts it down.

connorch/skills

Skill Claude CodeCodex

Compact the current conversation into a handoff document for another agent to pick up.

connorch/skills

Skill Claude CodeCodex

Controls claude-router, the local proxy that spreads Claude Code requests across subscription accounts. Use when Claude Code hits a rate limit, when asked which account is serving, or to bypass or force an account.

connorch/skills

Skill Claude CodeCodex

Use for any browser QA, verification, screenshot, or "check it in the browser" step when the T3 Code Browser panel tools (mcpt3-codepreview, such as previewstatus and previewsnapshot) are available. Drives the Browser panel the user watches live, in place of the repo's own browser automation.

connorch/skills

Skill Claude CodeCodex needs its repo

Use when the user asks for a motion graphic, animated explainer, or explainer video of a pull request, or of how something works ("make a video explaining X"). Renders it with Kit Langton's psychopomp engine and publishes the MP4 privately to files.wovn.org.

connorch/skills

Skill Codex

Rates how risky it is to merge a pull request or branch, on a Minimal / Low / Moderate / High / Critical scale, with the blockers that set the rating. The assessment runs in a subagent so it is independent of the thread that wrote the code. Use when asked to assess the merge risk of a PR or branch.