Skip to content

Commit e04d212

Browse files
GeneAIclaude
authored and
GeneAI
committed
chore: Release v4.4.0 - Production-Ready Agent Team System
Features: - Real LLM agent execution with Anthropic client integration - AskUserQuestion tool integration for form collection - Intent detection system for natural language agent creation - Skill-based invocation for agent teams (/release-prep, /test-coverage, etc.) - Security fix: Add usedforsecurity=False to MD5 hash calls 79 files changed across source, tests, VSCode extension, and Claude commands. Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
1 parent 64d50b1 commit e04d212

File tree

79 files changed

+10162
-355
lines changed

Some content is hidden

Large Commits have some content hidden by default. Use the searchbox below for content that may be hidden.

79 files changed

+10162
-355
lines changed

‎.claude/commands/create-agent.md‎

Lines changed: 77 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,77 @@
1+
# Create Custom Agent - Socratic Guide
2+
3+
You are helping the user create a custom AI agent for the Empathy Framework. Use the AskUserQuestion tool to gather requirements through a guided conversation.
4+
5+
## Step 1: Understand the Purpose
6+
7+
First, ask the user what they want their agent to do:
8+
9+
Use AskUserQuestion with:
10+
- Question: "What should this agent do?"
11+
- Header: "Purpose"
12+
- Options:
13+
- "Analyze code" - Review and analyze source code for issues or patterns
14+
- "Generate content" - Create tests, documentation, or other content
15+
- "Review & validate" - Check work for quality, security, or correctness
16+
- "Transform data" - Convert, migrate, or restructure code/data
17+
18+
## Step 2: Determine Specialization
19+
20+
Based on their answer, ask about specific focus:
21+
22+
Use AskUserQuestion with:
23+
- Question: "What specific area should this agent focus on?"
24+
- Header: "Focus Area"
25+
- Options vary by purpose (e.g., for "Analyze code"):
26+
- "Security vulnerabilities"
27+
- "Performance issues"
28+
- "Code quality & style"
29+
- "Architecture & design"
30+
31+
## Step 3: Select Model Tier
32+
33+
Use AskUserQuestion with:
34+
- Question: "What quality/cost balance do you need?"
35+
- Header: "Model Tier"
36+
- Options:
37+
- "Cheap (Recommended)" - Fast & low-cost ($0.001-0.01), good for simple analysis
38+
- "Capable" - Balanced ($0.01-0.05), good for most development tasks
39+
- "Premium" - Highest quality ($0.05-0.20), for complex reasoning
40+
41+
## Step 4: Define Success Criteria
42+
43+
Use AskUserQuestion with:
44+
- Question: "How will you measure success?"
45+
- Header: "Success"
46+
- Options:
47+
- "Issues found & reported" - Agent finds and documents problems
48+
- "Content generated" - Agent produces requested output
49+
- "Validation passed" - Agent confirms quality/correctness
50+
- "Recommendations provided" - Agent suggests improvements
51+
52+
## Step 5: Generate the Agent Spec
53+
54+
After gathering all answers, generate the agent specification:
55+
56+
```json
57+
{
58+
"name": "[Generated from purpose + focus]",
59+
"role": "[Description based on answers]",
60+
"tier": "[Selected tier]",
61+
"base_template": "generic",
62+
"success_criteria": "[Selected criteria]",
63+
"tools": []
64+
}
65+
```
66+
67+
Then tell the user:
68+
1. Show the generated spec in a code block
69+
2. Offer to save it: `empathy meta-workflow create-agent -q --name "X" --role "Y" --tier "Z" -o agent-spec.json`
70+
3. Explain how to use it in a team: `empathy meta-workflow create-team`
71+
72+
## Important
73+
74+
- Use AskUserQuestion for EACH step - don't ask multiple questions at once
75+
- Wait for user response before proceeding to next step
76+
- Keep the conversation focused and efficient
77+
- Generate a descriptive agent name based on their choices

‎.claude/commands/create-team.md‎

Lines changed: 95 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,95 @@
1+
# Create Custom Agent Team - Socratic Guide
2+
3+
You are helping the user create a custom AI agent team for their project using the Empathy Framework. Use the AskUserQuestion tool to gather requirements through a guided conversation.
4+
5+
## Step 1: Understand the Team's Mission
6+
7+
Use AskUserQuestion with:
8+
- Question: "What is the team's overall goal?"
9+
- Header: "Team Goal"
10+
- Options:
11+
- "Code quality pipeline" - Review, test, and improve code quality
12+
- "Release preparation" - Prepare code for production deployment
13+
- "Documentation sync" - Keep docs aligned with code
14+
- "Security audit" - Comprehensive security analysis
15+
16+
## Step 2: Determine Team Size
17+
18+
Use AskUserQuestion with:
19+
- Question: "How many agents should be on this team?"
20+
- Header: "Team Size"
21+
- Options:
22+
- "2 agents (Recommended)" - Simple workflow: analyze then act
23+
- "3 agents" - Standard: analyze, execute, validate
24+
- "4 agents" - Comprehensive: analyze, execute, validate, report
25+
- "5 agents" - Full pipeline with multiple specialists
26+
27+
## Step 3: Define Collaboration Pattern
28+
29+
Use AskUserQuestion with:
30+
- Question: "How should agents work together?"
31+
- Header: "Collaboration"
32+
- Options:
33+
- "Sequential (Recommended)" - Each agent waits for the previous one
34+
- "Parallel then merge" - Run analysis in parallel, then synthesize
35+
- "Pipeline" - Output of each feeds into the next
36+
37+
## Step 4: Select Agent Roles
38+
39+
Based on team size and goal, present role options. For a 3-agent team:
40+
41+
Use AskUserQuestion with multiSelect: true:
42+
- Question: "Which roles should your team have? (Select 3)"
43+
- Header: "Roles"
44+
- Options:
45+
- "Analyst" - Examines code/docs and identifies issues
46+
- "Generator" - Creates content (tests, docs, fixes)
47+
- "Reviewer" - Checks quality and correctness
48+
- "Validator" - Verifies results meet criteria
49+
50+
## Step 5: Cost Preference
51+
52+
Use AskUserQuestion with:
53+
- Question: "What's your cost preference for this team?"
54+
- Header: "Cost"
55+
- Options:
56+
- "Minimize cost (Recommended)" - Use cheap tier where possible ($0.03-0.10 per run)
57+
- "Balance cost/quality" - Mix of cheap and capable ($0.10-0.30 per run)
58+
- "Maximize quality" - Use capable/premium tiers ($0.30-0.60 per run)
59+
60+
## Step 6: Generate the Team Template
61+
62+
After gathering all answers, generate the team specification:
63+
64+
```json
65+
{
66+
"id": "[goal-based-id]",
67+
"name": "[Descriptive Team Name]",
68+
"description": "[Based on goal]",
69+
"collaboration_pattern": "[Selected pattern]",
70+
"agents": [
71+
{
72+
"role": "[Role 1]",
73+
"purpose": "[What this agent does]",
74+
"tier": "[Based on cost preference]",
75+
"base_template": "generic"
76+
}
77+
],
78+
"estimated_cost_range": {
79+
"min": 0.03,
80+
"max": 0.30
81+
}
82+
}
83+
```
84+
85+
Then tell the user:
86+
1. Show the generated template in a code block
87+
2. Save it to `.empathy/meta_workflows/templates/[id].json`
88+
3. Explain how to run it: `empathy meta-workflow run [id]`
89+
90+
## Important
91+
92+
- Use AskUserQuestion for EACH step - don't ask multiple questions at once
93+
- Wait for user response before proceeding to next step
94+
- For role selection, use multiSelect: true
95+
- Generate meaningful agent names and purposes based on the goal

‎.claude/commands/docs.md‎

Lines changed: 15 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -86,3 +86,18 @@ After running, provide:
8686
- Cost incurred
8787
- Links to generated docs
8888
- Suggestions for improvement
89+
90+
## Need Comprehensive Documentation Audit?
91+
92+
For automated documentation management with an AI agent team, use:
93+
94+
- `/manage-docs` - Full documentation gap analysis and sync verification
95+
96+
Or run directly:
97+
98+
```bash
99+
empathy meta-workflow run manage-docs --real --use-defaults
100+
```
101+
102+
This deploys specialized agents for Documentation Analyst, Documentation Reviewer,
103+
and Documentation Synthesizer to ensure docs stay in sync with code.

‎.claude/commands/feature-overview.md‎

Lines changed: 155 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,155 @@
1+
# Feature Overview Generator
2+
3+
Generate comprehensive technical documentation for code modules, suitable for architects, engineers, and content creators.
4+
5+
## Usage
6+
7+
```
8+
/feature-overview [path]
9+
```
10+
11+
**Arguments:**
12+
- `path` (optional): Directory or file to analyze. Defaults to current directory.
13+
14+
**Examples:**
15+
```
16+
/feature-overview src/empathy_os/meta_workflows/
17+
/feature-overview src/empathy_os/memory/
18+
/feature-overview .
19+
```
20+
21+
## What It Does
22+
23+
This command runs a 5-agent pipeline to analyze code and generate:
24+
25+
1. **Code Scanner** - Identifies modules, classes, functions, and their relationships
26+
2. **Insights Reporter** - Highlights architectural decisions and design patterns
27+
3. **Release Notes Writer** - Documents features in changelog-friendly format
28+
4. **Quality Checker** - Validates completeness and accuracy
29+
5. **Blog Content Creator** - Transforms insights into developer-friendly content
30+
31+
## Instructions for Claude
32+
33+
When the user invokes `/feature-overview`, follow these steps:
34+
35+
### Step 1: Determine Scope
36+
37+
If no path provided, ask:
38+
39+
Use AskUserQuestion with:
40+
- Question: "What would you like me to analyze?"
41+
- Header: "Scope"
42+
- Options:
43+
- "Current directory" - Analyze the entire project
44+
- "Specific module" - I'll ask which module
45+
- "Recent changes" - Focus on files changed since last commit
46+
47+
### Step 2: Identify Target Audience
48+
49+
Use AskUserQuestion with:
50+
- Question: "Who is the primary audience for this overview?"
51+
- Header: "Audience"
52+
- Options:
53+
- "Architects (Recommended)" - Focus on design patterns and architecture
54+
- "Engineers" - Focus on implementation details and APIs
55+
- "Technical Writers" - Focus on documentation and examples
56+
- "All" - Comprehensive coverage
57+
58+
### Step 3: Analyze the Code
59+
60+
Read the target files and analyze:
61+
- Module structure and organization
62+
- Key classes and their responsibilities
63+
- Design patterns used
64+
- Data flow and dependencies
65+
- Extension points and APIs
66+
67+
### Step 4: Generate the Overview
68+
69+
Create a comprehensive technical document with:
70+
71+
```markdown
72+
# Feature Overview: [Module Name]
73+
74+
**Generated:** [Date]
75+
**Modules Analyzed:** [Count]
76+
**Target Audience:** [Selected audience]
77+
78+
---
79+
80+
## Executive Summary
81+
82+
[2-3 sentence overview of what this code does]
83+
84+
## Architecture Overview
85+
86+
[Diagram or description of component relationships]
87+
88+
## Core Components
89+
90+
### [Component 1]
91+
- **Purpose:** [What it does]
92+
- **Location:** [file:line]
93+
- **Key Methods:** [List important methods]
94+
95+
### [Component 2]
96+
...
97+
98+
## Design Patterns
99+
100+
| Pattern | Location | Purpose |
101+
|---------|----------|---------|
102+
| [Pattern] | [file] | [Why used] |
103+
104+
## Data Flow
105+
106+
[Description of how data moves through the system]
107+
108+
## Extension Points
109+
110+
| Extension | How to Extend |
111+
|-----------|---------------|
112+
| [Point] | [Instructions] |
113+
114+
## Key Insights
115+
116+
1. [Architectural decision and rationale]
117+
2. [Notable pattern or approach]
118+
3. [Potential improvement or consideration]
119+
120+
## Blog-Ready Summary
121+
122+
> [2-3 paragraph summary suitable for a technical blog post]
123+
124+
---
125+
126+
*Generated by Empathy Framework Feature Overview*
127+
```
128+
129+
### Step 5: Offer Follow-ups
130+
131+
After generating the overview, ask:
132+
133+
Use AskUserQuestion with:
134+
- Question: "What would you like to do next?"
135+
- Header: "Next Steps"
136+
- Options:
137+
- "Save to file" - Save as markdown in docs/
138+
- "Generate blog post" - Create full blog post draft
139+
- "Analyze another module" - Run on different code
140+
- "Done" - Finish
141+
142+
## Quality Settings
143+
144+
This command uses the **maximize quality** tier configuration:
145+
- Insights Reporter: `capable` tier
146+
- Quality Checker: `premium` tier
147+
- Blog Content Creator: `premium` tier
148+
149+
Estimated cost: $0.40-0.80 per run with real LLM execution.
150+
151+
## Related Commands
152+
153+
- `/release-prep` - Prepare a full release with changelog
154+
- `/docs` - Generate API documentation
155+
- `/marketing` - Create marketing content from features

‎.claude/commands/manage-docs.md‎

Lines changed: 44 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,44 @@
1+
# /manage-docs - Documentation Management Agent Team
2+
3+
Invoke an AI agent team to ensure documentation stays in sync with code.
4+
5+
## What This Does
6+
7+
Creates and executes a team of specialized agents:
8+
- **Documentation Analyst** - Identifies missing docstrings and stale docs
9+
- **Documentation Reviewer** - Validates findings and removes false positives
10+
- **Documentation Synthesizer** - Creates prioritized action plan
11+
12+
## Usage
13+
14+
```
15+
/manage-docs
16+
```
17+
18+
## Instructions for Claude
19+
20+
When the user invokes /manage-docs, execute this workflow:
21+
22+
1. Run the documentation management meta-workflow with real execution:
23+
```bash
24+
python -m empathy_os.meta_workflows.cli_meta_workflows run manage-docs --real --use-defaults
25+
```
26+
27+
2. If ANTHROPIC_API_KEY is not set, fall back to mock execution:
28+
```bash
29+
python -m empathy_os.meta_workflows.cli_meta_workflows run manage-docs --use-defaults
30+
```
31+
32+
3. Present the results to the user with:
33+
- Missing docstrings count
34+
- Outdated documentation identified
35+
- README freshness status
36+
- Prioritized improvement suggestions
37+
38+
## Expected Output
39+
40+
The command will output:
41+
- Documentation gap analysis
42+
- Stale documentation detection
43+
- API documentation completeness
44+
- Prioritized action items for documentation improvements

0 commit comments

Comments
 (0)