notque/vexjoy-agent / skills/process/workflow
workflow
VerifiedStructured work: multi-phase tasks, feature builds, planning, objective loops, hill climbing.
references/comprehensive-review/references/wave-0-per-package.md
View on GitHub# Wave 0: Per-Package Deep Review
## Package Discovery Commands
| Language | Discovery Command | Agent Type | Min Package Size |
|----------|-------------------|------------|-----------------|
| Go | `find . -name "*.go" -path "*/internal/*" \| xargs dirname \| sort -u` | `general-purpose` + `programming` skill, `references/golang-general.md` | 1 file |
| Go (also) | `find . -name "*.go" -not -path "*/internal/*" -not -path "*/vendor/*" \| xargs dirname \| sort -u` | `general-purpose` + `programming` skill, `references/golang-general.md` | 1 file |
| Python | `find . -name "__init__.py" \| xargs dirname \| sort -u` | `general-purpose` + `programming` skill, `references/python-general.md` | 1 file |
| TypeScript | `find . -name "index.ts" -o -name "index.tsx" \| xargs dirname \| sort -u` | `general-purpose` + `frontend` skill, `references/typescript-frontend.md` | 1 file |
Full discovery commands (excluding vendor/venv/git):
```bash
# Go packages (internal/ and top-level)
find . -name "*.go" -not -path "*/vendor/*" -not -path "*/.git/*" | xargs dirname | sort -u
# Python packages
find . -name "__init__.py" -not -path "*/venv/*" -not -path "*/.git/*" | xargs dirname | sort -u
# TypeScript modules (directories with index.ts/tsx)
find . -name "index.ts" -o -name "index.tsx" | grep -v node_modules | xargs dirname | sort -u
```
## Agent Selection by Language
| Language | Worker |
|----------|-------|
| Go | `general-purpose` + `programming` skill, `references/golang-general.md` |
| Python | `general-purpose` + `programming` skill, `references/python-general.md` |
| TypeScript | `general-purpose` + `frontend` skill, `references/typescript-frontend.md` |
| Mixed | Use language of majority files in that package |
Review agents run on `opus` (the `work` role in `workflow-helpers.js`); pass the owner's model instead when the owner names one. Never use haiku or sonnet for code review.
## Per-Package Agent Dispatch Prompt
Dispatch up to 10 agents per message. If more than 10 packages are discovered, use multiple batches:
- Batch 1: packages 1-10 (dispatch in ONE message)
- Batch 2: packages 11-20 (dispatch in ONE message after batch 1 completes)
- Continue until all packages are covered
Each per-package worker is a fresh `general-purpose` agent. Its prompt starts by calling the Skill tool with the row's skill and reading `~/.claude/skills/<skill>/<reference>`, then continues with this prompt:
```
PER-PACKAGE DEEP REVIEW — Wave 0
PACKAGE: [package path]
LANGUAGE: [Go/Python/TypeScript]
MCP TOOL DISCOVERY (do this FIRST, before reading package files):
- Use ToolSearch to check for available MCP tools that can enhance your analysis:
a. Run ToolSearch("gopls") — if this is a Go package, loads type-aware analysis
tools (go_file_context, go_diagnostics, go_symbol_references, etc.)
b. Run ToolSearch("context7") — loads library documentation lookup tools
- If gopls tools are available AND LANGUAGE is Go:
* Use go_file_context after reading each .go file for intra-package dependency context
* Use go_diagnostics on the package to detect build/analysis errors
* Use go_symbol_references to check for unused or misused exported symbols
- If Context7 tools are available:
* Use resolve-library-id + query-docs for unfamiliar library APIs in this package
INSTRUCTIONS:
1. Read the CLAUDE.md file(s) in this repository first
2. Run MCP TOOL DISCOVERY steps above
3. Read EVERY file in this package directory: [package path]/
4. Understand the package's purpose, internal APIs, and relationships
5. Review ALL code for issues — you have full package context
6. Use MCP tools (gopls, Context7) as you review for type-aware precision
7. Focus on issues that require understanding the WHOLE package:
- Internal API misuse between files in this package
- Inconsistent error handling patterns within the package
- Missing or redundant functionality
- Package-level design issues (cohesion, coupling)
- Test coverage relative to package complexity
8. Use structured output format with severity classification
9. Include file:line references for every finding
CONTEXT: This is Wave 0 of a comprehensive review. Your per-package findings
will be passed to 20 cross-cutting review agents in Waves 1 and 2. Focus on
issues that require full package context to detect — cross-cutting agents will
handle file-level and project-level concerns.
OUTPUT FORMAT:
### PACKAGE: [package path]
**Files reviewed**: [list all files read]
**Package purpose**: [1-sentence summary]
**Package health**: [HEALTHY | MINOR_ISSUES | NEEDS_ATTENTION | CRITICAL]
Findings:
### [CRITICAL|HIGH|MEDIUM|LOW]: [One-line summary]
**File**: `path/to/file:LINE`
**Issue**: [Description]
**Impact**: [Why this matters]
**Fix**: [Concrete code fix]
**Requires package context**: [Why a single-file reviewer would miss this]
---
```
## Wave 0 Aggregate Output Format
After all per-package agents complete, build this summary for Wave 1+2 context injection:
```markdown
## Wave 0 Per-Package Findings Summary (for Wave 1+2 context)
### Packages Reviewed: [N]
### Packages Healthy: [N] | Minor Issues: [N] | Needs Attention: [N] | Critical: [N]
### Per-Package Results
#### [package/path1] — [HEALTHY|MINOR_ISSUES|NEEDS_ATTENTION|CRITICAL]
- Purpose: [1-sentence]
- Files: [N]
- Findings: [N] (CRITICAL: N, HIGH: N, MEDIUM: N, LOW: N)
- Key issues:
- [SEVERITY]: [summary] at [file:line]
- ...
#### [package/path2] — [STATUS]
- ...
### Cross-Package Patterns Detected
- [Pattern 1]: Seen in packages [list] — suggests systemic issue
- [Pattern 2]: ...
### Wave 0 Hotspots (packages with most findings)
1. [package] — [N findings, N critical]
2. [package] — [N findings, N high]
3. ...
```
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| planning-and-task-breakdownaddyosmani/agent-skillsBreaks work into ordered tasks. Use when you have a spec or clear requirements and need to break work into implementable tasks. Use when a task feels too large to start, when you need to estimate scope, or when parallel work is possible. | Stars: 119k | Updated | Claude CodeCodex | Verified | Install |
| screenpipethedotmack/claude-memSearch screen and audio history recorded by Screenpipe and optionally save selected context into claude-mem. Use when the user asks to recall desktop activity, a meeting, or something seen in another app through Screenpipe. | Stars: 118k | Updated | Claude CodeCodex | Verified | Install |
| contact-widgetnexu-io/open-designSelf-contained floating chat widget with welcome screen, social links, meeting button, and message input. Single HTML file, zero dependencies. | Stars: 108k | Updated | Claude CodeCodex | Verified | Install |
| meeting-notesnexu-io/open-designMeeting notes page — title bar with attendees, agenda checklist, decisions block, action items table with owners + dates, and a "next meeting" footer. Use when the brief mentions "meeting notes", "minutes", "1:1 notes", "all-hands recap", or "会议纪要". | Stars: 108k | Updated | Claude CodeCodex | Verified | Install |