cmtraceopen-code-review

Use when reviewing any cmtraceopen change - a diff, branch, or PR - or when asked whether cmtraceopen work is merge-ready. Runs the review against the repo's own gates instead of generic criteria.

Jump to install

Repository: adamgell/cmtraceopen
Last source activity: August 8, 2026 at 21:40
Detected SKILL.md language: English
Stars: 281
Forks: 15

Install options

The review-first prompt is selected by default. You can switch to a direct command or download a local copy.

Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.

A direct command skips the review prompt. Inspect the source before running it.

npx skills add https://github.com/adamgell/cmtraceopen --skill cmtraceopen-code-review

The command stays on one line. Scroll horizontally to inspect it before copying.

Prefer a local copy? Download the files currently available to SkillsMP.

Review the source files

Read SKILL.md and any companion files shown by SkillsMP before deciding whether to install.

SKILL.md

Source instructions · Read-only preview

Back to search Back to installation

CMTrace Open — Code Review

Thin wrapper. The canonical review contract is .Clairvoyance/staff/code-review-charter.md — read it FIRST and follow its load order and review recipe exactly (contract layer, then adversarial layer, then mechanical layer).

The charter routes to everything else: the routing indexes (.Clairvoyance/library.md, repo-root library.md), the four reducer ADRs (docs/architecture/decisions/), the reducer review checklist (docs/superpowers/plans/2026-08-07-reducer-framework-v1.md), the reviewer role charters (reducer-contract, reducer-adversary, reducer-integration), and the specialist context (repo-root soul.md, memory.md).

The deliverable is the charter's gate-state report: findings ranked most-severe first, named gate states (CI, CodeRabbit approved_at_head via the coderabbit-review-loop skill's state script, contract conformance), and rejected feedback with reasoning. Merging is the repository owner's action; the review ends at the report.

Delegating a review to Hermes

Launch rules, each learned from a failed launch (2026-08-08):

Everything the review needs must exist ON MAIN and on disk before launching - an unmerged skill or charter branch means Hermes finds nothing and wanders into stale refs.

Launch from the repo directory so relative paths resolve, and state in the prompt that all charter files are on main.

The prompt must say, verbatim: "read branch files ONLY via git show origin/BRANCH:PATH - never git checkout." A --yolo session in the shared root checkout once checked out the PR branch, which yanked the charter and skills out from under every other session. Verify git branch --show-current is main after any Hermes session ends.

Do not pass -m: Hermes routes per his own tiering (his direct API keys are unset; gateway model names are not enumerable from the CLI). If a GPT tier is explicitly required, Adam's pick is luna, never sol.

Redirect output straight to a log file, never through a pipe (| tail buffers everything and hides a silent death). hermes --cli also buffers stdout until exit when non-interactive, so the desktop UI is the live view - monitor the observable outcome (the PR comment, the process), not the log.

After launching, verify the process exists (pgrep) before reporting it running, and arm a watcher that fires on success AND on death.

Working launch shape:

cd /Users/Adam.Gell/repo/cmtraceopen && hermes --cli --yolo \ --skills cmtraceopen-code-review \ -z "<task: target PR; charter on main; git show only, never checkout; read-only; one report comment>" \ > /tmp/hermes-review.log 2>&1 &