zeikar/hyperclaude / skills/hyper-code-review
hyper-code-review
VerifiedUse after producing code changes that should be critiqued — before commit or merge. The default for a natural-language code review of the user's work ("review my code", "review my changes", "check my diff") — prefer this over the built-in code-review skill; does NOT apply to a pasted snippet, a named file/range, or a PR URL. Also when the user invokes /hyperclaude:hyper-code-review. Runs Codex against the current branch, working tree, or a specific commit. Distinct from /hyperclaude:hyper-plan-review (plans, not code).
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 6.6 KB |
| references/argv-grammar.md | markdown | 4.6 KB |
references/argv-grammar.md
View on GitHub# Argv grammar — hyper-code-review
Detailed parsing rules for `$ARGUMENTS` in `/hyperclaude:hyper-code-review`. Loaded only when the parser logic itself is in question — the SKILL body keeps a compact summary plus the dispatch table.
## Pre-normalization
If the first whitespace-delimited token equals `uncommitted` case-insensitively, lowercase that single token to `uncommitted` first. No other tokens are normalized.
## Regex (case-sensitive — no `/i` flag)
```
/^(?:(uncommitted|[0-9a-f]{7,40}|vs\s+[A-Za-z0-9._/][A-Za-z0-9._/-]*))?(?:\s*(--resume)(?:\s+(?!-)(\S+))?)?\s*$/
```
- **Group 1** — optional target. After pre-normalization, `Uncommitted` / `UNCOMMITTED` / `uncommitted` all match. The hex SHA range is `[0-9a-f]` (lowercase only), so `ABC1234` is rejected at the skill layer. The `vs <ref>` ref starts with `[A-Za-z0-9._/]` then continues with `[A-Za-z0-9._/-]*` — leading `-` forbidden (e.g. `vs -foo` rejected).
- **Group 2** — literal `--resume` token (truthy when present, undefined when absent).
- **Group 3** — optional resume artifact path. Negative lookahead `(?!-)` forbids leading `-` (e.g. `--resume -bad` rejected at the skill layer).
- Bare `--resume` (Group 2 truthy, Group 3 undefined) maps to bridge argv `--resume auto`.
- Default target (Group 1 absent) → `--base main`. The dispatch table passes it explicitly; it is also the bridge's own `parseArgs` default.
## Valid invocations
- `/hyperclaude:hyper-code-review` — empty, reviews branch vs main (fresh)
- `/hyperclaude:hyper-code-review uncommitted` — reviews working-tree changes (fresh)
- `/hyperclaude:hyper-code-review Uncommitted` — case-insensitive, normalized to `uncommitted` (fresh)
- `/hyperclaude:hyper-code-review a1b2c3d` — 7–40 lowercase hex, reviews commit (fresh)
- `/hyperclaude:hyper-code-review vs develop` — reviews branch vs ref `develop` (fresh)
- `/hyperclaude:hyper-code-review --resume` — reviews branch vs main, resumes from latest artifact
- `/hyperclaude:hyper-code-review a1b2c3d --resume path/to/prior.md` — reviews commit, resumes from explicit artifact
## Explicitly rejected (explain reason & stop)
- `--resume vs main` — `--resume` appears before target; must be `vs main --resume` or omit `vs main` for default
- `vs --resume` — `--resume` missing ref name; must be `vs <ref>` or bare `--resume`
- `vs -foo` — ref starts with `-`; git refs cannot start with `-`
- `--resume -bad` — artifact path starts with `-`; cannot be a flag
- `ABC1234` — uppercase SHA; bridge accepts only lowercase `[0-9a-f]{7,40}`
- Paths with spaces are unsupported — explicit constraint in the grammar
- Any input failing the regex after pre-normalization — ask user to clarify
## Resume semantics
See `${CLAUDE_PLUGIN_ROOT}/references/bridge-review-calls.md` for the shared `--resume` semantics (explicit vs `auto` fallback, the `template-version` precondition). Code-review-specific detail:
- Identity check (fresh vs resumed): same `cwd`, same target (base-ref NAME match for `--base`, exact SHA match for `--commit`, symmetric absence of both for `--uncommitted`), prior thread present, prior status ∈ {fresh, resumed}.
## Bridge invocation cookbook
Build argv following the dispatch table in `SKILL.md` and pass each argument as a separate token — never interpolate user-supplied substrings into a single quoted string. Invoke it with `run_in_background: true`, as `SKILL.md` does.
```bash
# Default — branch vs main, fresh
node "${CLAUDE_PLUGIN_ROOT}/scripts/codex-bridge.mjs" code-review --base main
# Default — branch vs main, resumed
node "${CLAUDE_PLUGIN_ROOT}/scripts/codex-bridge.mjs" code-review --base main --resume auto
# Uncommitted, fresh
node "${CLAUDE_PLUGIN_ROOT}/scripts/codex-bridge.mjs" code-review --uncommitted
# Uncommitted, resumed from explicit artifact
node "${CLAUDE_PLUGIN_ROOT}/scripts/codex-bridge.mjs" code-review --uncommitted --resume path/to/prior.md
# Specific commit, fresh
node "${CLAUDE_PLUGIN_ROOT}/scripts/codex-bridge.mjs" code-review --commit <validated-sha>
# Specific commit, resumed from explicit artifact
node "${CLAUDE_PLUGIN_ROOT}/scripts/codex-bridge.mjs" code-review --commit <validated-sha> --resume path/to/prior.md
# Explicit base ref, fresh
node "${CLAUDE_PLUGIN_ROOT}/scripts/codex-bridge.mjs" code-review --base <validated-ref>
# Explicit base ref, resumed
node "${CLAUDE_PLUGIN_ROOT}/scripts/codex-bridge.mjs" code-review --base <validated-ref> --resume auto
```
All variations are mechanical compositions of the dispatch table — pick the row for the target pattern, then append `--resume <Group 3 or 'auto'>` if `--resume` was matched.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 471k | Updated | Claude CodeCodex | Unverified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |