Issue Investigation & Planning Skill
SkillFiles & storageInvestigate a GitHub issue and produce a structured implementation plan. Fetches issue details, explores the codebase, identifies affected layers (UI, BLoC, Repository, Client), and outputs a step-by-step plan with files to modify, testing strategy, and risk assessment. Works for bugs, features, and tasks. Invoke with /plan.
Available today. Use it from your connected AI after setup.
No other account needed.
Connect ahel once, and every AI you use reads what you have installed.
Then ask your AI: use the Issue Investigation & Planning Skill skill
What this skill tells your AI
The instructions your AI receives, as published by divinevideo/divine-mobile in .agents/skills/plan/SKILL.md and read by ahel’s review.
Purpose
Investigate a GitHub issue thoroughly and produce a structured, actionable implementation plan. The plan follows the project's layered architecture (UI -> BLoC -> Repository -> Client) and includes files to modify, implementation steps, testing strategy, and risk assessment.
Workflow
Phase 1: Fetch Issue
Retrieve the full issue context using gh CLI and GraphQL.
# Get issue details
gh issue view <number> --json title,body,comments,labels,state,assignees
# Get issue type, hierarchy, and sub-issues
gh api graphql -f query='{ repository(owner: "divinevideo", name: "divine-mobile") {
issue(number: <NUMBER>) {
id
title
body
issueType { id name }
subIssues(first: 20) { nodes { number title state } }
parentIssue { number title }
}
} }'
Determine the issue type from:
- Title prefix:
fix:= Bug,feat:= Feature,task:= Task - GraphQL
issueTypefield - Labels (e.g.,
buglabel)
Extract key information:
- Bug: Steps to reproduce, actual vs expected result, environment, evidence
- Feature: Requirements, acceptance criteria, user stories
- Task: Description, scope, implementation notes
Phase 2: Explore Codebase
Search the codebase to understand the relevant code. Approach varies by type.
For Bugs
- Search for keywords from the bug description (error messages, screen names, widget names) using
Grep - Identify the screen/page where the bug manifests using
Globonlib/screens/andlib/widgets/ - Trace the code path from UI through BLoC to Repository to Client
- Check for related tests using
Globontest/ - If Nostr-related, use
mcp__nostr__read_kindormcp__nostr__read_nipfor protocol context
For Features
- Search for similar existing features using
GrepandGlob - Identify which architectural layers are needed (new BLoC? New repository? New package?)
- Check if related models/repositories already exist in
mobile/packages/ - Look at the router (
mobile/lib/router/) for navigation integration - If Nostr-related, use
mcp__nostr__*tools to understand relevant NIPs and event kinds
For Tasks
- Search for the specific code area mentioned in the task using
Grep - Understand the current implementation state
- Check for existing tests that need updating
- Identify related files that might be affected
Phase 3: Analyze
For Bugs
- Identify the root cause by reading the relevant source files
- Determine which layer(s) contain the bug (UI? BLoC? Repository? Client?)
- Assess regression risk (what else could break?)
- Check if existing tests should have caught this
For Features
- Determine which layers need new code vs modifications
- Design the data flow: Client -> Repository -> BLoC -> UI
- Identify new models, events, states needed
- Consider integration with existing features (router, theme, existing BLoCs)
- Assess if any new packages are needed in
mobile/packages/
For Tasks
- Scope the change precisely
- Identify all files that need modification
- Determine if the task has cascading effects
Phase 4: Plan
Construct the implementation plan following the project's layered architecture. Always order implementation steps bottom-up:
- Data layer changes (Client packages in
mobile/packages/) - Repository layer changes (Repository packages in
mobile/packages/) - Business logic changes (BLoCs in
mobile/lib/blocs/) - Presentation layer changes (Screens/Widgets in
mobile/lib/screens/ormobile/lib/widgets/) - Router changes (if navigation is affected)
- Test plan (mirroring the implementation layers)
Phase 5: Output
Present the plan using the output template below.
Output Template
## Plan: [Issue Title] (#[number])
**Type**: Bug | Feature | Task
**Issue**: [URL]
**Complexity**: Low | Medium | High
---
### Issue Summary
[1-2 sentence summary in your own words after investigation]
### [Bug Only] Root Cause
- **Layer**: [UI | BLoC | Repository | Client]
- **File**: [exact file path with line numbers]
- **Cause**: [clear explanation of why this happens]
- **Evidence**: [code snippet or logic trace]
### [Feature Only] Architecture Design
- **New packages needed**: [list or "None"]
- **Layers affected**: [which of UI, BLoC, Repository, Client]
- **Data flow**: [Client -> Repository -> BLoC -> UI description]
- **Nostr events**: [relevant kinds/NIPs if applicable]
### Affected Files
| File | Action | Description |
|------|--------|-------------|
| `path/to/file.dart` | Modify / Create / Delete | What changes |
### Implementation Steps
Steps ordered bottom-up (data layer first, UI last):
1. **[Layer]: [Brief description]**
- File: `path/to/file.dart`
- Changes: [specific changes]
- Why: [rationale]
2. ...
### Testing Strategy
| Layer | Test File | What to Test |
|-------|-----------|-------------|
| [Layer] | `test/path/to/test.dart` | [specific test cases] |
### Risks and Considerations
- [Risk 1 and mitigation]
- [Risk 2 and mitigation]
Tool Reference
| Need | Tool | Example |
|---|---|---|
| Fetch issue | gh issue view | gh issue view 142 --json title,body,comments,labels |
| Issue hierarchy | gh api graphql | Sub-issues, parent issues, issue type |
| Find files by name | Glob | Glob("**/*video_feed*") |
| Search code content | Grep | Grep("VideoFeedBloc") |
| Read file details | Read | Read("mobile/lib/blocs/video_feed/video_feed_bloc.dart") |
| Check code health | mcp__dart__analyze_files | Analyze affected packages |
| Find Dart symbols | mcp__dart__resolve_workspace_symbol | Look up class/method names |
| Nostr protocol | mcp__nostr__read_nip | Check NIP specifications |
| Nostr event kinds | mcp__nostr__read_kind | Understand event structure |
Complexity Guidelines
| Complexity | Criteria |
|---|---|
| Low | Single layer, 1-3 files, no new models or packages |
| Medium | 2-3 layers, 4-10 files, may need new BLoC events/states |
| High | All layers, 10+ files, new packages, new models, Nostr protocol changes |
Architecture Reference
- Layer order: UI -> BLoC -> Repository -> Client
- Implementation order: Bottom-up (Client first, UI last)
- Packages:
mobile/packages/ - BLoCs:
mobile/lib/blocs/ - Screens:
mobile/lib/screens/ - Widgets:
mobile/lib/widgets/ - Router:
mobile/lib/router/ - Tests mirror:
mobile/test/mirrorsmobile/lib/ - State management: BLoC for new features, Riverpod for legacy
- Theme: Dark mode only, use
VineThemeconstants - Nostr IDs: Never truncate, always full 64-char hex
Signals
- GitHub stars
- 265
- Forks
- 55
- Last commit
- Sep 2026
Advanced
- Catalog kind
- skill
- Gateway key
plan-divinevideo- Source
- github.com/divinevideo/divine-mobile