oliver-kriska/claude-elixir-phoenix / targets/opencode/skills/phx-review
phx-review
VerifiedReview changed Elixir/Phoenix code read-only. Check requirements, cite evidence, deduplicate findings, and return a severity-based verdict.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 4.1 KB |
| references/agent-spawning.md | markdown | 1.5 KB |
| references/blocker-handling.md | markdown | 2.6 KB |
| references/conventions.md | markdown | 1.9 KB |
| references/example-review.md | markdown | 2.5 KB |
| references/requirements-detection.md | markdown | 1.3 KB |
| references/review-template.md | markdown | 3.4 KB |
references/agent-spawning.md
View on GitHub# OpenCode Review Execution Reference
`/phx-review` works without separately installed custom agents. Native OpenCode
subagents are an optional performance optimization, not a correctness dependency.
## Concern Selection
| Concern | Select when |
|---|---|
| Elixir/Phoenix | Always |
| Security | Auth, session, password, token, upload, or input code changed |
| Testing | Tests changed, public behavior changed, or regression coverage is absent |
| Ecto/LiveView | Relevant schemas, queries, migrations, LiveViews, or components changed |
| Oban | Worker or queue code changed |
| Deployment | Dockerfile, release, `fly.toml`, or runtime configuration changed |
| Requirements | A plan, specification, or issue is available |
## Parallel Mode
When native subagent tooling is available, delegate independent concerns to
generic read-only workers. Give each worker the same base ref, changed-file list,
requirements context, and instruction to return evidence-backed findings with
`path:line` citations. Keep one worker per concern and deduplicate in the parent.
When subagents are unavailable, expensive, or unnecessary, review the same
concerns sequentially in the current session. Never treat a sequential run as a
failed review and never require plugin-root agent definitions or Claude task APIs.
## Output Contract
Return findings to the current session. Do not write files unless the user asks
for a persisted report. The review remains read-only and ends after the verdict.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| receiving-code-reviewobra/superpowersUse when receiving code review feedback, before implementing suggestions, especially if feedback seems unclear or technically questionable - requires technical rigor and verification, not performative agreement or blind implementation | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| requesting-code-reviewobra/superpowersUse when completing tasks, implementing major features, or before merging to verify work meets requirements | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| using-git-worktreesobra/superpowersUse when starting feature work that needs isolation from current workspace or before executing implementation plans - ensures an isolated workspace exists via native tools or git worktree fallback | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| verification-before-completionobra/superpowersUse when about to claim work is complete, fixed, or passing, before committing or creating PRs - requires running verification commands and confirming output before making any success claims; evidence before assertions always | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |