Skip to content

Commit 1265067

Browse files
gracefullightgracefullight
committed
feat: introduce new skills and resources for developer workflow
1 parent e65757a commit 1265067

34 files changed

Lines changed: 3317 additions & 654 deletions

‎.agent/brain/.gitkeep‎

Whitespace-only changes.

‎.agent/skills/_shared/context-loading.md‎

Lines changed: 52 additions & 36 deletions
Original file line numberDiff line numberDiff line change
@@ -31,60 +31,76 @@ This saves context window and prevents confusion from irrelevant information.
3131

3232
### Backend Agent
3333

34-
| Task Type | Required Resources |
35-
|-----------|-------------------|
36-
| CRUD API creation | snippets.md (route, schema, model, test) |
34+
| Task Type | Required Resources |
35+
| ----------------------------- | ------------------------------------------- |
36+
| CRUD API creation | snippets.md (route, schema, model, test) |
3737
| Authentication implementation | snippets.md (JWT, password) + tech-stack.md |
38-
| DB migration | snippets.md (migration) |
39-
| Performance optimization | examples.md (N+1 example) |
40-
| Existing code modification | examples.md + Serena MCP |
38+
| DB migration | snippets.md (migration) |
39+
| Performance optimization | examples.md (N+1 example) |
40+
| Existing code modification | examples.md + Serena MCP |
4141

4242
### Frontend Agent
4343

44-
| Task Type | Required Resources |
45-
|-----------|-------------------|
46-
| Component creation | snippets.md (component, test) + component-template.tsx |
47-
| Form implementation | snippets.md (form + Zod) |
48-
| API integration | snippets.md (TanStack Query) |
49-
| Styling | tailwind-rules.md |
50-
| Page layout | snippets.md (grid) + examples.md |
44+
| Task Type | Required Resources |
45+
| ------------------- | ------------------------------------------------------ |
46+
| Component creation | snippets.md (component, test) + component-template.tsx |
47+
| Form implementation | snippets.md (form + Zod) |
48+
| API integration | snippets.md (TanStack Query) |
49+
| Styling | tailwind-rules.md |
50+
| Page layout | snippets.md (grid) + examples.md |
5151

5252
### Mobile Agent
5353

54-
| Task Type | Required Resources |
55-
|-----------|-------------------|
56-
| Screen creation | snippets.md (screen, provider) + screen-template.dart |
57-
| API integration | snippets.md (repository, Dio) |
58-
| Navigation | snippets.md (GoRouter) |
59-
| Offline features | examples.md (offline example) |
60-
| State management | snippets.md (Riverpod) |
54+
| Task Type | Required Resources |
55+
| ---------------- | ----------------------------------------------------- |
56+
| Screen creation | snippets.md (screen, provider) + screen-template.dart |
57+
| API integration | snippets.md (repository, Dio) |
58+
| Navigation | snippets.md (GoRouter) |
59+
| Offline features | examples.md (offline example) |
60+
| State management | snippets.md (Riverpod) |
6161

6262
### Debug Agent
6363

64-
| Task Type | Required Resources |
65-
|-----------|-------------------|
66-
| Frontend bug | common-patterns.md (Frontend section) |
67-
| Backend bug | common-patterns.md (Backend section) |
68-
| Mobile bug | common-patterns.md (Mobile section) |
64+
| Task Type | Required Resources |
65+
| --------------- | ----------------------------------------------------------------- |
66+
| Frontend bug | common-patterns.md (Frontend section) |
67+
| Backend bug | common-patterns.md (Backend section) |
68+
| Mobile bug | common-patterns.md (Mobile section) |
6969
| Performance bug | common-patterns.md (Performance section) + debugging-checklist.md |
70-
| Security bug | common-patterns.md (Security section) |
70+
| Security bug | common-patterns.md (Security section) |
7171

7272
### QA Agent
7373

74-
| Task Type | Required Resources |
75-
|-----------|-------------------|
76-
| Security review | checklist.md (Security section) |
77-
| Performance review | checklist.md (Performance section) |
74+
| Task Type | Required Resources |
75+
| -------------------- | ------------------------------------ |
76+
| Security review | checklist.md (Security section) |
77+
| Performance review | checklist.md (Performance section) |
7878
| Accessibility review | checklist.md (Accessibility section) |
79-
| Full audit | checklist.md (full) + self-check.md |
79+
| Full audit | checklist.md (full) + self-check.md |
80+
81+
### Developer Workflow Expert
82+
83+
| Task Type | Required Resources |
84+
| --------------------------- | ------------------------------------------------------------- |
85+
| API Workflow Setup | resources/api-workflows.md + resources/validation-pipeline.md |
86+
| Database Migration Workflow | resources/database-patterns.md |
87+
| Release Coordination | resources/release-coordination.md |
88+
| Troubleshooting | resources/troubleshooting.md |
89+
90+
### Terraform Infra Engineer
91+
92+
| Task Type | Required Resources |
93+
| --------------------------- | ------------------------------------------------------------------------ |
94+
| Infrastructure Provisioning | resources/multi-cloud-examples.md + resources/policy-testing-examples.md |
95+
| Cost Analysis | resources/cost-optimization.md |
8096

8197
### PM Agent
8298

83-
| Task Type | Required Resources |
84-
|-----------|-------------------|
85-
| New project planning | examples.md + task-template.json + api-contracts/template.md |
86-
| Feature addition planning | examples.md + Serena MCP (understand existing structure) |
87-
| Refactoring planning | Serena MCP only |
99+
| Task Type | Required Resources |
100+
| ------------------------- | ------------------------------------------------------------ |
101+
| New project planning | examples.md + task-template.json + api-contracts/template.md |
102+
| Feature addition planning | examples.md + Serena MCP (understand existing structure) |
103+
| Refactoring planning | Serena MCP only |
88104

89105
---
90106

‎.agent/skills/_shared/multi-review-protocol.md‎

Lines changed: 75 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,75 @@
1+
# Multi-Review Protocol
2+
3+
## Core Principle
4+
**Every task is reviewed multiple times from different perspectives.**
5+
6+
---
7+
8+
## Review Types Guide
9+
10+
### 1. Completeness Review (Step 2)
11+
- **Question**: "Is anything missing?"
12+
- **Check**: Map requirements to plan items
13+
- **Pass Condition**: All requirements reflected in plan
14+
15+
### 2. Meta Review (Step 3)
16+
- **Question**: "Was the review done properly?"
17+
- **Check**: Self-verify previous review was sufficient
18+
- **Pass Condition**: No review gaps confirmed
19+
20+
### 3. Simplicity Review (Step 4)
21+
- **Question**: "Is this over-engineered?"
22+
- **Check**: Question necessity of each component
23+
- **Remove**: "Might need later", speculative features
24+
25+
### 4. Alignment Review (Step 6)
26+
- **Question**: "Did we build what was requested?"
27+
- **Check**: Compare plan vs implementation
28+
- **Pass Condition**: 1:1 mapping confirmed
29+
30+
### 5. Safety Review (Step 7)
31+
- **Question**: "Is there anything dangerous?"
32+
- **Check**: OWASP Top 10, potential bugs
33+
- **Tools**: npm audit, bandit, lighthouse
34+
- **Pass Condition**: Zero CRITICAL/HIGH issues
35+
36+
### 6. Regression Review (Step 8)
37+
- **Question**: "Did improvements break anything?"
38+
- **Check**: Existing tests pass, existing features work
39+
- **Pass Condition**: No regressions
40+
41+
### 7. Reusability Review (Step 10)
42+
- **Question**: "Can we leverage existing code?"
43+
- **Check**: Similar functions/components exist
44+
- **Action**: Integrate if reusable
45+
46+
### 8. Consistency Review (Step 12)
47+
- **Question**: "Is everything harmonious?"
48+
- **Check**: Naming, style, architecture consistency
49+
- **Pass Condition**: Aligns with existing codebase
50+
51+
### 9. Quality Review (Step 14)
52+
- **Question**: "Does it meet quality standards?"
53+
- **Check**: lint, types, coverage, complexity
54+
- **Pass Condition**: All quality metrics pass
55+
56+
### 10. Cascade Impact Review (Step 16)
57+
- **Question**: "Did we break anything elsewhere?"
58+
- **Check**: Use find_referencing_symbols for impact scope
59+
- **Pass Condition**: No cascade impact or handled
60+
61+
### 11. Final Review (Step 17)
62+
- **Question**: "Is this ready to deploy?"
63+
- **Check**: Complete checklist final verification
64+
- **Pass Condition**: User final approval
65+
66+
---
67+
68+
## Failure Recovery
69+
70+
| Review | Return Point on Failure |
71+
|--------|------------------------|
72+
| Step 2-4 | Step 1 (Revise plan) |
73+
| Step 6-8 | Step 5 (Fix implementation) |
74+
| Step 10-13 | Step 9 (Restart refinement) |
75+
| Step 14-17 | Appropriate phase based on failure |

‎.agent/skills/_shared/phase-gates.md‎

Lines changed: 110 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,110 @@
1+
# Phase Gate Definitions
2+
3+
Each phase must pass its gate before proceeding to the next.
4+
5+
---
6+
7+
## PLAN_GATE
8+
9+
**Owner**: PM Agent
10+
**Trigger**: After Steps 1-4
11+
12+
### Criteria
13+
- [ ] Plan documented with acceptance criteria
14+
- [ ] Assumptions explicitly listed
15+
- [ ] Alternatives considered for architecture decisions (min 2)
16+
- [ ] Over-engineering review completed
17+
- [ ] User confirmation received
18+
19+
### Auto-pass Conditions
20+
- Difficulty: Simple
21+
- Existing pattern match
22+
- User explicitly skips
23+
24+
### Failure Action
25+
Revise plan, do not proceed to IMPL
26+
27+
---
28+
29+
## IMPL_GATE
30+
31+
**Owner**: Implementation Agent
32+
**Trigger**: After Step 5
33+
34+
### Criteria
35+
- [ ] Code compiles/builds successfully
36+
- [ ] Tests pass
37+
- [ ] Only planned files modified
38+
- [ ] No unrequested features added
39+
- [ ] Diff reviewed for scope creep
40+
41+
### Auto-pass Conditions
42+
- All tests green
43+
- Diff < 200 lines
44+
- No new dependencies
45+
46+
### Failure Action
47+
Fix issues, re-run implementation
48+
49+
---
50+
51+
## VERIFY_GATE
52+
53+
**Owner**: QA Agent
54+
**Trigger**: After Steps 6-8
55+
56+
### Criteria
57+
- [ ] Implementation matches requirements
58+
- [ ] Zero CRITICAL issues
59+
- [ ] Zero HIGH issues
60+
- [ ] Improvements validated (no regressions)
61+
62+
### Blockers
63+
- Any CRITICAL or HIGH issue
64+
65+
### Failure Action
66+
Return to IMPL with findings
67+
68+
---
69+
70+
## REFINE_GATE
71+
72+
**Owner**: Implementation + Debug Agents
73+
**Trigger**: After Steps 9-13
74+
75+
### Criteria
76+
- [ ] No files > 500 lines (or justified)
77+
- [ ] No functions > 50 lines (or justified)
78+
- [ ] Integration opportunities captured
79+
- [ ] Side effects verified
80+
- [ ] Unused code cleaned
81+
82+
### Skip Conditions
83+
- Simple tasks < 50 lines total change
84+
- User explicitly skips
85+
86+
### Failure Action
87+
Address issues, re-verify
88+
89+
---
90+
91+
## SHIP_GATE
92+
93+
**Owner**: QA Agent
94+
**Trigger**: After Steps 14-17
95+
96+
### Criteria
97+
- [ ] Lint passes
98+
- [ ] Type check passes
99+
- [ ] Test coverage >= 80%
100+
- [ ] UX flows verified
101+
- [ ] No hardcoded secrets
102+
- [ ] Migrations safe
103+
- [ ] Related issues addressed
104+
- [ ] Deployment checklist complete
105+
106+
### Final Approval
107+
User must confirm
108+
109+
### Failure Action
110+
Return to appropriate phase based on failure type

‎.agent/skills/_shared/quality-principles.md‎

Lines changed: 82 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,82 @@
1+
# Oh-My-AG Quality Principles
2+
3+
4 principles to follow across all development phases.
4+
5+
---
6+
7+
## 1. Think Before Coding
8+
9+
**Core**: Surface assumptions, ask when uncertain, present alternatives.
10+
11+
### Application
12+
- List assumptions explicitly
13+
- Present 2+ alternatives for architecture decisions
14+
- Stop and clarify when confused
15+
- Push back on unclear requirements
16+
17+
### Anti-patterns
18+
- Starting without understanding
19+
- Silently choosing one interpretation
20+
- Hiding confusion
21+
22+
---
23+
24+
## 2. Simplicity First
25+
26+
**Core**: Minimum code that solves the problem. Nothing speculative.
27+
28+
### Application
29+
- Implement exactly what's specified
30+
- No "while I'm here" additions
31+
- Skip abstractions for single-use code
32+
- Rewrite if 200 lines could be 50
33+
34+
### Anti-patterns
35+
- Feature creep
36+
- "Might need later" code
37+
- Unrequested flexibility
38+
39+
---
40+
41+
## 3. Surgical Changes
42+
43+
**Core**: Touch only what you must. Clean up only your own mess.
44+
45+
### Application
46+
- Modify only task-related files
47+
- Match existing style
48+
- Remove only orphans YOUR changes created
49+
- Flag unrelated issues without fixing
50+
51+
### Anti-patterns
52+
- Drive-by refactoring
53+
- Improving adjacent code
54+
- Cleaning pre-existing debt
55+
56+
---
57+
58+
## 4. Verify Until Done
59+
60+
**Core**: Define success criteria. Loop until verified.
61+
62+
### Application
63+
- Transform tasks into verifiable goals
64+
- "Add validation" → "Write failing tests, then pass them"
65+
- "Fix bug" → "Create reproducing test, then pass it"
66+
- Run automated verification at each checkpoint
67+
68+
### Anti-patterns
69+
- Declaring victory without testing
70+
- Assuming correctness
71+
- Skipping verification steps
72+
73+
---
74+
75+
## Quick Reference
76+
77+
| Principle | One-liner | Question to Ask |
78+
|-----------|-----------|-----------------|
79+
| Think Before Coding | "What could go wrong?" | "Did I list my assumptions?" |
80+
| Simplicity First | "Do we need this?" | "Would a senior call this overcomplicated?" |
81+
| Surgical Changes | "Is this in scope?" | "Does every line trace to the request?" |
82+
| Verify Until Done | "How do I prove it works?" | "What's my success criteria?" |

‎.agent/skills/_version.json‎

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -1,3 +1,3 @@
11
{
2-
"version": "1.20.0"
2+
"version": "1.25.0"
33
}

‎.agent/skills/debug-agent/SKILL.md‎

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -22,7 +22,7 @@ description: Bug diagnosis and fixing specialist - analyzes errors, identifies r
2222
3. Minimal fix: change only what's necessary
2323
4. Every fix gets a regression test
2424
5. Search for similar patterns elsewhere after fixing
25-
6. Document in `.gemini/antigravity/brain/bugs/`
25+
6. Document in `.agent/brain/bugs/`
2626

2727
## How to Execute
2828
Follow `resources/execution-protocol.md` step by step.

0 commit comments

Comments
 (0)