Menü
Run parallel code reviews with both Claude and Codex, then synthesize findings
Mit Codex oder Claude installieren Kopieren Sie diesen Prompt, fügen Sie ihn in Codex, Claude oder einen anderen Assistant ein und lassen Sie die Skill-Seite prüfen und installieren.
Basierend auf der SOC-Berufsklassifikation
How to invoke the Codex CLI (`codex`) from the shell. Read this BEFORE shelling out to `codex` for any nontrivial task — it covers calling the real binary (not the Superconductor wrapper, which hangs) and capturing output with `-o` so the answer doesn't get buried in the agent-session stream.
Interact with Google Docs & Drive — read, edit, search, comment, share, create docs, list/pull/push content. Use whenever JP wants to read or modify a Google Doc or Drive file (fetch a doc's text, leave a comment, find a file, etc.).
Batch-process Asana tickets — bugs and small / well-specified features — tagged jpa-bugfixes-today into PRs. Spawns a subagent per ticket in parallel worktrees with Codex plan-review and diff-review.
Interact with Asana — tasks, My Tasks, projects, sections, tags, users, workspaces. Use whenever JP wants to read or modify Asana data (look up or edit a task, list My Tasks, comment, find a project, etc.).
Query Better Stack Telemetry logs — errors, warnings, full-text search, request traces, tail, stats. Use whenever JP wants to look at logs or investigate the context around a failure (what happened around an error, logs for a request id, recent errors, etc.).
Interact with Sentry — issues, errors, events, traces, releases, projects, dashboards. Use whenever JP wants to read or modify Sentry data (triage an issue, look up an error, inspect a trace, manage a release, etc.).
| name | review-multi |
| description | Run parallel code reviews with both Claude and Codex, then synthesize findings |
| allowed-tools | ["Task","Agent","Read","Edit","Bash","Grep","Glob"] |
This skill runs two code reviews in parallel: one via the installed
review-claude adapter and one via the installed review-codex adapter. Then
it presents both sets of findings.
/review-multi or $review-multi - run reviews, then discuss findings/review-multi fix or $review-multi fix - run reviews, then fix issues
autonomously/review-multi base main or $review-multi base main - review changes
against a base branch/review-multi mini or $review-multi mini - faster reviewModes can be combined: /review-multi mini fix base main.
Parse arguments from: $ARGUMENTS in Claude, or the user's prompt after
$review-multi in Codex.
Determine mode:
Read the skill definitions from the sibling skill directories:
../review-claude/SKILL.md../review-codex/SKILL.mdThese define the review procedures for each reviewer. Follow their execution instructions, but with these modifications:
Launch both Claude and Codex reviews in parallel. Pass along the base branch argument if present.
Launch both Claude and Codex reviews in parallel, but pass mini to both
review adapters.
Collaborative mode: After showing all outputs, provide your own synthesis. Deduplicate the findings, and summarize the points to address. Offer to help address issues found. Do not make fixes yourself in this mode. Do this by default.
Pass-through mode: Display both review outputs directly. Do not add commentary, analysis, or suggestions.
Fix mode: After reviewing both outputs, make executive decisions and fix issues autonomously. Use your own judgment on what's worth fixing.
One warning: Your judgment tends to be too lenient on type system issues, fail-fast violations, and backwards-compatibility hacks. When either reviewer flags these, lean toward fixing them rather than dismissing them.
After fixing, briefly summarize what you changed and why. If you skipped any flagged issues, explain your reasoning.
The default mode is collaborative — if no mode argument is provided, synthesize the findings and offer to help with fixes, but don't fix autonomously.