mirror of
https://github.com/catlog22/Claude-Code-Workflow.git
synced 2026-03-03 15:43:11 +08:00
feat: add multi-mode workflow planning skill with session management and task generation
This commit is contained in:
catlog22
89
.claude/skills/team-perf-opt/subagents/discuss-subagent.md
Normal file
89
.claude/skills/team-perf-opt/subagents/discuss-subagent.md
Normal file
@@ -0,0 +1,89 @@
|
||||
# Discuss Subagent
|
||||
|
||||
Multi-perspective discussion for evaluating optimization strategies and reviewing code change quality. Used by strategist (DISCUSS-OPT) and reviewer (DISCUSS-REVIEW) when complex trade-offs require multi-angle analysis.
|
||||
|
||||
## Design Rationale
|
||||
|
||||
Complex optimization decisions (e.g., choosing between algorithmic change vs caching layer) and nuanced code review findings (e.g., evaluating whether a side effect is acceptable) benefit from structured multi-perspective analysis. This subagent provides that analysis inline without spawning additional team members.
|
||||
|
||||
## Invocation
|
||||
|
||||
Called by strategist, reviewer after their primary analysis when complexity warrants multi-perspective evaluation:
|
||||
|
||||
```
|
||||
Task({
|
||||
subagent_type: "cli-discuss-agent",
|
||||
run_in_background: false,
|
||||
description: "Discuss <round-id>: <topic> for performance optimization",
|
||||
prompt: `Conduct a multi-perspective discussion on the following topic.
|
||||
|
||||
Round: <round-id>
|
||||
Topic: <discussion-topic>
|
||||
Session: <session-folder>
|
||||
|
||||
Context:
|
||||
<relevant-context-from-calling-role>
|
||||
|
||||
Perspectives to consider:
|
||||
- Performance impact: Will this actually improve the target metric?
|
||||
- Risk assessment: What could go wrong? Side effects? Regressions?
|
||||
- Maintainability: Is the optimized code still understandable and maintainable?
|
||||
- Alternative approaches: Are there simpler or safer ways to achieve the same goal?
|
||||
|
||||
Evaluate trade-offs and provide a structured recommendation with:
|
||||
- Consensus verdict: proceed / revise / escalate
|
||||
- Confidence level: high / medium / low
|
||||
- Key trade-offs identified
|
||||
- Recommended approach with rationale
|
||||
- Dissenting perspectives (if any)`
|
||||
})
|
||||
```
|
||||
|
||||
## Round Configuration
|
||||
|
||||
| Round | Artifact | Parameters | Calling Role |
|
||||
|-------|----------|------------|-------------|
|
||||
| DISCUSS-OPT | <session>/discussions/DISCUSS-OPT.md | Optimization strategy trade-offs | strategist |
|
||||
| DISCUSS-REVIEW | <session>/discussions/DISCUSS-REVIEW.md | Code review finding validation | reviewer |
|
||||
|
||||
## Integration with Calling Role
|
||||
|
||||
The calling role is responsible for:
|
||||
|
||||
1. **Before calling**: Complete primary analysis, identify the specific trade-off or finding needing discussion
|
||||
2. **Calling**: Invoke subagent with round ID, topic, and relevant context
|
||||
3. **After calling**:
|
||||
|
||||
| Result | Action |
|
||||
|--------|--------|
|
||||
| consensus_reached (proceed) | Incorporate recommendation into output, continue |
|
||||
| consensus_reached (revise) | Adjust findings/strategy based on discussion insights |
|
||||
| consensus_blocked (HIGH) | Report to coordinator via message with severity |
|
||||
| consensus_blocked (MEDIUM) | Include in output with recommendation for revision |
|
||||
| consensus_blocked (LOW) | Note in output, proceed with original assessment |
|
||||
|
||||
## Output Schema
|
||||
|
||||
```json
|
||||
{
|
||||
"round_id": "<DISCUSS-OPT|DISCUSS-REVIEW>",
|
||||
"topic": "<discussion-topic>",
|
||||
"verdict": "<proceed|revise|escalate>",
|
||||
"confidence": "<high|medium|low>",
|
||||
"trade_offs": [
|
||||
{ "dimension": "<performance|risk|maintainability>", "pro": "<benefit>", "con": "<cost>" }
|
||||
],
|
||||
"recommendation": "<recommended-approach>",
|
||||
"rationale": "<reasoning>",
|
||||
"dissenting_views": ["<alternative-perspective>"]
|
||||
}
|
||||
```
|
||||
|
||||
## Error Handling
|
||||
|
||||
| Scenario | Resolution |
|
||||
|----------|------------|
|
||||
| Single perspective analysis fails | Continue with partial perspectives |
|
||||
| All analyses fail | Return basic recommendation from calling role's primary analysis |
|
||||
| Artifact not found | Return error immediately |
|
||||
| Discussion inconclusive | Return "revise" verdict with low confidence |
|
||||
108
.claude/skills/team-perf-opt/subagents/explore-subagent.md
Normal file
108
.claude/skills/team-perf-opt/subagents/explore-subagent.md
Normal file
@@ -0,0 +1,108 @@
|
||||
# Explore Subagent
|
||||
|
||||
Shared codebase exploration for discovering performance-critical code paths, module structures, and optimization opportunities. Results are cached to avoid redundant exploration across profiler and optimizer roles.
|
||||
|
||||
## Design Rationale
|
||||
|
||||
Codebase exploration is a read-only operation shared between profiler (mapping bottlenecks) and optimizer (understanding implementation context). Caching explorations avoids redundant work when optimizer re-explores paths the profiler already mapped.
|
||||
|
||||
## Invocation
|
||||
|
||||
Called by profiler, optimizer after needing codebase context for performance analysis or implementation:
|
||||
|
||||
```
|
||||
Task({
|
||||
subagent_type: "cli-explore-agent",
|
||||
run_in_background: false,
|
||||
description: "Explore codebase for performance-critical paths in <target-scope>",
|
||||
prompt: `Explore the codebase to identify performance-critical code paths.
|
||||
|
||||
Target scope: <target-scope>
|
||||
Session: <session-folder>
|
||||
Focus: <exploration-focus>
|
||||
|
||||
Tasks:
|
||||
1. Map the module structure and entry points within scope
|
||||
2. Identify hot code paths (frequently called functions, critical loops)
|
||||
3. Find performance-relevant patterns (caching, lazy loading, async, pooling)
|
||||
4. Note any existing performance optimizations or benchmark harnesses
|
||||
5. List key files with their roles in the performance-critical path
|
||||
|
||||
Output a structured exploration report with:
|
||||
- Module map (key files and their relationships)
|
||||
- Hot path analysis (call chains, loop nests, recursive patterns)
|
||||
- Existing optimization patterns found
|
||||
- Performance-relevant configuration (caching, pooling, batching settings)
|
||||
- Recommended investigation targets for profiling`
|
||||
})
|
||||
```
|
||||
|
||||
## Cache Mechanism
|
||||
|
||||
### Cache Index Schema
|
||||
|
||||
`<session-folder>/explorations/cache-index.json`:
|
||||
|
||||
```json
|
||||
{
|
||||
"entries": [
|
||||
{
|
||||
"key": "<scope-hash>",
|
||||
"scope": "<target-scope>",
|
||||
"focus": "<exploration-focus>",
|
||||
"timestamp": "<ISO-8601>",
|
||||
"result_file": "<hash>.md"
|
||||
}
|
||||
]
|
||||
}
|
||||
```
|
||||
|
||||
### Cache Lookup Rules
|
||||
|
||||
| Condition | Action |
|
||||
|-----------|--------|
|
||||
| Exact scope+focus match exists | Return cached result from <hash>.md |
|
||||
| No match | Execute subagent, cache result to <hash>.md, update index |
|
||||
| Cache file missing but index has entry | Remove stale entry, re-execute |
|
||||
| Cache older than current session | Use cached (explorations are stable within session) |
|
||||
|
||||
## Integration with Calling Role
|
||||
|
||||
The calling role is responsible for:
|
||||
|
||||
1. **Before calling**: Determine target scope and exploration focus
|
||||
2. **Calling**: Check cache first, invoke subagent only on cache miss
|
||||
3. **After calling**:
|
||||
|
||||
| Result | Action |
|
||||
|--------|--------|
|
||||
| Exploration successful | Use findings to inform profiling/implementation |
|
||||
| Exploration partial | Use available findings, note gaps |
|
||||
| Exploration failed | Proceed without exploration context, use direct file reading |
|
||||
|
||||
## Output Schema
|
||||
|
||||
```json
|
||||
{
|
||||
"scope": "<target-scope>",
|
||||
"module_map": [
|
||||
{ "file": "<path>", "role": "<description>", "hot_path": true }
|
||||
],
|
||||
"hot_paths": [
|
||||
{ "chain": "<call-chain>", "frequency": "<high|medium|low>", "files": ["<path>"] }
|
||||
],
|
||||
"existing_optimizations": [
|
||||
{ "type": "<pattern>", "location": "<file:line>", "description": "<what>" }
|
||||
],
|
||||
"investigation_targets": ["<file-or-pattern>"]
|
||||
}
|
||||
```
|
||||
|
||||
## Error Handling
|
||||
|
||||
| Scenario | Resolution |
|
||||
|----------|------------|
|
||||
| Single exploration angle fails | Continue with partial results |
|
||||
| All exploration fails | Return basic result from direct file listing |
|
||||
| Target scope not found | Return error immediately |
|
||||
| Cache corrupt | Clear cache-index.json, re-execute |
|
||||
Reference in New Issue
Block a user