@leeovery/claude-technical-workflows 2.0.16 → 2.0.18
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/README.md +41 -24
- package/commands/link-dependencies.md +9 -9
- package/commands/start-discussion.md +6 -313
- package/commands/start-feature.md +80 -0
- package/commands/start-implementation.md +7 -162
- package/commands/start-planning.md +7 -86
- package/commands/start-research.md +7 -13
- package/commands/start-specification.md +7 -81
- package/commands/workflow:start-discussion.md +339 -0
- package/commands/workflow:start-implementation.md +187 -0
- package/commands/workflow:start-planning.md +111 -0
- package/commands/workflow:start-research.md +38 -0
- package/commands/workflow:start-review.md +112 -0
- package/commands/workflow:start-specification.md +106 -0
- package/package.json +1 -1
- package/skills/technical-discussion/SKILL.md +13 -8
- package/skills/technical-implementation/SKILL.md +23 -11
- package/skills/technical-planning/SKILL.md +17 -13
- package/skills/technical-planning/references/formal-planning.md +97 -5
- package/skills/technical-research/SKILL.md +13 -9
- package/skills/technical-review/SKILL.md +26 -9
- package/skills/technical-review/references/review-checklist.md +4 -2
- package/skills/technical-specification/SKILL.md +16 -9
- package/skills/technical-specification/references/specification-guide.md +8 -9
|
@@ -0,0 +1,112 @@
|
|
|
1
|
+
---
|
|
2
|
+
description: Start a review session from an existing plan and implementation. Discovers available plans, validates implementation exists, and invokes the technical-review skill.
|
|
3
|
+
---
|
|
4
|
+
|
|
5
|
+
Invoke the **technical-review** skill for this conversation.
|
|
6
|
+
|
|
7
|
+
## Workflow Context
|
|
8
|
+
|
|
9
|
+
This is **Phase 6** of the six-phase workflow:
|
|
10
|
+
|
|
11
|
+
| Phase | Focus | You |
|
|
12
|
+
|-------|-------|-----|
|
|
13
|
+
| 1. Research | EXPLORE - ideas, feasibility, market, business | |
|
|
14
|
+
| 2. Discussion | WHAT and WHY - decisions, architecture, edge cases | |
|
|
15
|
+
| 3. Specification | REFINE - validate into standalone spec | |
|
|
16
|
+
| 4. Planning | HOW - phases, tasks, acceptance criteria | |
|
|
17
|
+
| 5. Implementation | DOING - tests first, then code | |
|
|
18
|
+
| **6. Review** | VALIDATING - check work against artifacts | HERE |
|
|
19
|
+
|
|
20
|
+
**Stay in your lane**: Verify that every plan task was implemented, tested adequately, and meets quality standards. Don't fix code - identify problems. You're reviewing, not building.
|
|
21
|
+
|
|
22
|
+
---
|
|
23
|
+
|
|
24
|
+
## Instructions
|
|
25
|
+
|
|
26
|
+
Follow these steps EXACTLY as written. Do not skip steps or combine them.
|
|
27
|
+
|
|
28
|
+
Before beginning, discover existing work and gather necessary information.
|
|
29
|
+
|
|
30
|
+
## Important
|
|
31
|
+
|
|
32
|
+
Use simple, individual commands. Never combine multiple operations into bash loops or one-liners. Execute commands one at a time.
|
|
33
|
+
|
|
34
|
+
## Step 1: Discover Existing Plans
|
|
35
|
+
|
|
36
|
+
Scan the codebase for plans:
|
|
37
|
+
|
|
38
|
+
1. **Find plans**: Look in `docs/workflow/planning/`
|
|
39
|
+
- Run `ls docs/workflow/planning/` to list plan files
|
|
40
|
+
- Each file is named `{topic}.md`
|
|
41
|
+
|
|
42
|
+
2. **Check plan format**: For each plan file
|
|
43
|
+
- Run `head -10 docs/workflow/planning/{topic}.md` to read the frontmatter
|
|
44
|
+
- Note the `format:` field
|
|
45
|
+
- Do NOT use bash loops - run separate `head` commands for each topic
|
|
46
|
+
|
|
47
|
+
## Step 2: Check Prerequisites
|
|
48
|
+
|
|
49
|
+
**If no plans exist:**
|
|
50
|
+
|
|
51
|
+
```
|
|
52
|
+
No plans found in docs/workflow/planning/
|
|
53
|
+
|
|
54
|
+
The review phase requires a completed implementation based on a plan. Please run /workflow:start-planning first to create a plan, then /workflow:start-implementation to build it.
|
|
55
|
+
```
|
|
56
|
+
|
|
57
|
+
Stop here and wait for the user to acknowledge.
|
|
58
|
+
|
|
59
|
+
## Step 3: Present Options to User
|
|
60
|
+
|
|
61
|
+
Show what you found:
|
|
62
|
+
|
|
63
|
+
```
|
|
64
|
+
Plans found:
|
|
65
|
+
{topic-1} (format: {format})
|
|
66
|
+
{topic-2} (format: {format})
|
|
67
|
+
|
|
68
|
+
Which plan would you like to review the implementation for?
|
|
69
|
+
```
|
|
70
|
+
|
|
71
|
+
**Auto-select:** If exactly one plan exists, automatically select it and proceed to Step 4. Inform the user which plan was selected. Do not ask for confirmation.
|
|
72
|
+
|
|
73
|
+
## Step 4: Identify Implementation Scope
|
|
74
|
+
|
|
75
|
+
Determine what code to review:
|
|
76
|
+
|
|
77
|
+
1. **Check git status** - See what files have changed
|
|
78
|
+
2. **Ask user** if unclear:
|
|
79
|
+
- "What code should I review? (all changes, specific directories, or specific files)"
|
|
80
|
+
|
|
81
|
+
## Step 5: Locate Specification (Optional)
|
|
82
|
+
|
|
83
|
+
Check if a specification exists:
|
|
84
|
+
|
|
85
|
+
1. **Look for specification**: Check `docs/workflow/specification/{topic}.md`
|
|
86
|
+
2. **If exists**: Note the path for context
|
|
87
|
+
3. **If missing**: Proceed without - the plan is the primary review artifact
|
|
88
|
+
|
|
89
|
+
## Step 6: Invoke Review Skill
|
|
90
|
+
|
|
91
|
+
Pass to the technical-review skill:
|
|
92
|
+
- Plan: `docs/workflow/planning/{topic}.md`
|
|
93
|
+
- Format: (from frontmatter)
|
|
94
|
+
- Specification: `docs/workflow/specification/{topic}.md` (if exists)
|
|
95
|
+
- Implementation scope: (files/directories to review)
|
|
96
|
+
|
|
97
|
+
**Example handoff:**
|
|
98
|
+
```
|
|
99
|
+
Review session for: {topic}
|
|
100
|
+
Plan: docs/workflow/planning/{topic}.md
|
|
101
|
+
Format: {format}
|
|
102
|
+
Specification: docs/workflow/specification/{topic}.md (or: Not available)
|
|
103
|
+
Scope: {implementation scope}
|
|
104
|
+
|
|
105
|
+
Begin review using the technical-review skill.
|
|
106
|
+
```
|
|
107
|
+
|
|
108
|
+
## Notes
|
|
109
|
+
|
|
110
|
+
- Ask questions clearly and wait for responses before proceeding
|
|
111
|
+
- Review produces feedback (approve, request changes, or comments) - it does NOT fix code
|
|
112
|
+
- Verify ALL plan tasks, don't sample
|
|
@@ -0,0 +1,106 @@
|
|
|
1
|
+
---
|
|
2
|
+
description: Start a specification session from an existing discussion. Discovers available discussions, checks for existing specifications, and invokes the technical-specification skill.
|
|
3
|
+
---
|
|
4
|
+
|
|
5
|
+
Invoke the **technical-specification** skill for this conversation.
|
|
6
|
+
|
|
7
|
+
## Workflow Context
|
|
8
|
+
|
|
9
|
+
This is **Phase 3** of the six-phase workflow:
|
|
10
|
+
|
|
11
|
+
| Phase | Focus | You |
|
|
12
|
+
|-------|-------|-----|
|
|
13
|
+
| 1. Research | EXPLORE - ideas, feasibility, market, business | |
|
|
14
|
+
| 2. Discussion | WHAT and WHY - decisions, architecture, edge cases | |
|
|
15
|
+
| **3. Specification** | REFINE - validate into standalone spec | ◀ HERE |
|
|
16
|
+
| 4. Planning | HOW - phases, tasks, acceptance criteria | |
|
|
17
|
+
| 5. Implementation | DOING - tests first, then code | |
|
|
18
|
+
| 6. Review | VALIDATING - check work against artifacts | |
|
|
19
|
+
|
|
20
|
+
**Stay in your lane**: Validate and refine discussion content into a standalone specification. Don't jump to planning, phases, tasks, or code. The specification is the "line in the sand" - everything after this has hard dependencies on it.
|
|
21
|
+
|
|
22
|
+
---
|
|
23
|
+
|
|
24
|
+
## Instructions
|
|
25
|
+
|
|
26
|
+
Follow these steps EXACTLY as written. Do not skip steps or combine them. Present output using the EXACT format shown in examples - do not simplify or alter the formatting.
|
|
27
|
+
|
|
28
|
+
Before beginning, discover existing work and gather necessary information.
|
|
29
|
+
|
|
30
|
+
## Important
|
|
31
|
+
|
|
32
|
+
Use simple, individual commands. Never combine multiple operations into bash loops or one-liners. Execute commands one at a time.
|
|
33
|
+
|
|
34
|
+
## Step 1: Discover Existing Work
|
|
35
|
+
|
|
36
|
+
Scan the codebase for discussions and specifications:
|
|
37
|
+
|
|
38
|
+
1. **Find discussions**: Look in `docs/workflow/discussion/`
|
|
39
|
+
- Run `ls docs/workflow/discussion/` to list discussion files
|
|
40
|
+
- Each file is named `{topic}.md`
|
|
41
|
+
|
|
42
|
+
2. **Check discussion status**: For each discussion file
|
|
43
|
+
- Run `head -20 docs/workflow/discussion/{topic}.md` to read the frontmatter and extract the `status:` field
|
|
44
|
+
- Do NOT use bash loops - run separate `head` commands for each topic
|
|
45
|
+
|
|
46
|
+
3. **Check for existing specifications**: Look in `docs/workflow/specification/`
|
|
47
|
+
- Identify discussions that don't have corresponding specifications
|
|
48
|
+
|
|
49
|
+
## Step 2: Check Prerequisites
|
|
50
|
+
|
|
51
|
+
**If no discussions exist:**
|
|
52
|
+
|
|
53
|
+
```
|
|
54
|
+
⚠️ No discussions found in docs/workflow/discussion/
|
|
55
|
+
|
|
56
|
+
The specification phase requires a completed discussion. Please run /workflow:start-discussion first to document the technical decisions, edge cases, and rationale before creating a specification.
|
|
57
|
+
```
|
|
58
|
+
|
|
59
|
+
Stop here and wait for the user to acknowledge.
|
|
60
|
+
|
|
61
|
+
## Step 3: Present Options to User
|
|
62
|
+
|
|
63
|
+
Show what you found using a list like below:
|
|
64
|
+
|
|
65
|
+
```
|
|
66
|
+
📂 Discussions found:
|
|
67
|
+
✅ {topic-1} - Concluded - ready for specification
|
|
68
|
+
⚠️ {topic-2} - Exploring - not ready for specification
|
|
69
|
+
✅ {topic-3} - Concluded - specification exists
|
|
70
|
+
|
|
71
|
+
Which discussion would you like to create a specification for?
|
|
72
|
+
```
|
|
73
|
+
|
|
74
|
+
**Important:** Only concluded discussions should proceed to specification. If a discussion is still exploring, advise the user to complete the discussion phase first.
|
|
75
|
+
|
|
76
|
+
Ask: **Which discussion would you like to specify?**
|
|
77
|
+
|
|
78
|
+
## Step 4: Gather Additional Context
|
|
79
|
+
|
|
80
|
+
Ask:
|
|
81
|
+
- Any additional context or priorities to consider?
|
|
82
|
+
- Any constraints or changes since the discussion concluded?
|
|
83
|
+
- Are there any existing partial plans or related documentation I should review?
|
|
84
|
+
|
|
85
|
+
## Step 5: Invoke Specification Skill
|
|
86
|
+
|
|
87
|
+
Pass to the technical-specification skill:
|
|
88
|
+
- Source: `docs/workflow/discussion/{topic}.md`
|
|
89
|
+
- Output: `docs/workflow/specification/{topic}.md`
|
|
90
|
+
- Additional context gathered
|
|
91
|
+
|
|
92
|
+
**Example handoff:**
|
|
93
|
+
```
|
|
94
|
+
Specification session for: {topic}
|
|
95
|
+
Source: docs/workflow/discussion/{topic}.md
|
|
96
|
+
Output: docs/workflow/specification/{topic}.md
|
|
97
|
+
|
|
98
|
+
Begin specification using the technical-specification skill.
|
|
99
|
+
Reference: specification-guide.md
|
|
100
|
+
```
|
|
101
|
+
|
|
102
|
+
## Notes
|
|
103
|
+
|
|
104
|
+
- Ask questions clearly and wait for responses before proceeding
|
|
105
|
+
- The specification phase validates and refines discussion content into a standalone document
|
|
106
|
+
- Commit the specification files frequently during the session
|
package/package.json
CHANGED
|
@@ -7,16 +7,21 @@ description: "Document technical discussions as expert architect and meeting ass
|
|
|
7
7
|
|
|
8
8
|
Act as **expert software architect** participating in discussions AND **documentation assistant** capturing them. Do both simultaneously. Engage deeply while documenting for planning teams.
|
|
9
9
|
|
|
10
|
-
##
|
|
10
|
+
## Purpose in the Workflow
|
|
11
11
|
|
|
12
|
-
|
|
13
|
-
|
|
14
|
-
|
|
15
|
-
4. **Planning** (after): HOW - phases, tasks, acceptance criteria
|
|
16
|
-
5. **Implementation** (after): DOING - tests first, then code
|
|
17
|
-
6. **Review** (final): VALIDATING - check work against artifacts
|
|
12
|
+
This skill can be used:
|
|
13
|
+
- **Sequentially** (Phase 2): After research to debate and document decisions
|
|
14
|
+
- **Standalone** (Contract entry): To document technical decisions from any source
|
|
18
15
|
|
|
19
|
-
|
|
16
|
+
Either way: Capture decisions, rationale, competing approaches, and edge cases.
|
|
17
|
+
|
|
18
|
+
### What This Skill Needs
|
|
19
|
+
|
|
20
|
+
- **Topic** (required) - What technical area to discuss/document
|
|
21
|
+
- **Context** (optional) - Prior research, constraints, existing decisions
|
|
22
|
+
- **Questions to explore** (optional) - Specific architectural questions to address
|
|
23
|
+
|
|
24
|
+
**If missing:** Will ask user what topic they want to discuss.
|
|
20
25
|
|
|
21
26
|
## What to Capture
|
|
22
27
|
|
|
@@ -9,16 +9,23 @@ Act as **expert senior developer** who builds quality software through disciplin
|
|
|
9
9
|
|
|
10
10
|
Execute plans through strict TDD. Write tests first, then code to pass them.
|
|
11
11
|
|
|
12
|
-
##
|
|
12
|
+
## Purpose in the Workflow
|
|
13
13
|
|
|
14
|
-
|
|
15
|
-
|
|
16
|
-
|
|
17
|
-
4. **Planning** (previous): HOW - phases, tasks, acceptance criteria
|
|
18
|
-
5. **Implementation** (YOU): DOING - tests first, then code
|
|
19
|
-
6. **Review** (next): VALIDATING - check work against artifacts
|
|
14
|
+
This skill can be used:
|
|
15
|
+
- **Sequentially** (Phase 5): To execute a plan created by technical-planning
|
|
16
|
+
- **Standalone** (Contract entry): To execute any plan that follows plan-format conventions
|
|
20
17
|
|
|
21
|
-
|
|
18
|
+
Either way: Execute via strict TDD - tests first, implementation second.
|
|
19
|
+
|
|
20
|
+
### What This Skill Needs
|
|
21
|
+
|
|
22
|
+
- **Plan content** (required) - Phases, tasks, and acceptance criteria to execute
|
|
23
|
+
- **Plan format** (required) - How to parse tasks (from plan frontmatter)
|
|
24
|
+
- **Specification content** (optional) - For context when task rationale is unclear
|
|
25
|
+
- **Environment setup** (optional) - First-time setup instructions
|
|
26
|
+
- **Scope** (optional) - Specific phase/task to work on
|
|
27
|
+
|
|
28
|
+
**If missing:** Will ask user for plan location. If no specification, plan becomes sole authority.
|
|
22
29
|
|
|
23
30
|
## Hard Rules
|
|
24
31
|
|
|
@@ -51,9 +58,10 @@ Complete ALL setup steps before proceeding to implementation work.
|
|
|
51
58
|
|
|
52
59
|
See **[environment-setup.md](references/environment-setup.md)** for details.
|
|
53
60
|
|
|
54
|
-
2. **Read the plan** from `docs/workflow/planning/{topic}.md`
|
|
61
|
+
2. **Read the plan** from the provided location (typically `docs/workflow/planning/{topic}.md`)
|
|
55
62
|
- Check the `format` field in frontmatter
|
|
56
63
|
- Load the output adapter: `skills/technical-planning/references/output-{format}.md`
|
|
64
|
+
- If no format field, ask user which format the plan uses
|
|
57
65
|
- Follow the **Implementation** section for how to read tasks and update progress
|
|
58
66
|
|
|
59
67
|
3. **Read the TDD workflow** - Load **[tdd-workflow.md](references/tdd-workflow.md)** before writing any code. This is mandatory.
|
|
@@ -85,14 +93,18 @@ Keep user informed of progress:
|
|
|
85
93
|
|
|
86
94
|
## When to Reference Specification
|
|
87
95
|
|
|
88
|
-
Check the specification
|
|
96
|
+
Check the specification when:
|
|
89
97
|
|
|
90
98
|
- Task rationale is unclear
|
|
91
99
|
- Multiple valid approaches exist
|
|
92
100
|
- Edge case handling not specified in plan
|
|
93
101
|
- You need the "why" behind a decision
|
|
94
102
|
|
|
95
|
-
|
|
103
|
+
**Location**: Specification should be linked in the plan file (check frontmatter or plan header). Ask user if not found.
|
|
104
|
+
|
|
105
|
+
The specification (if available) is the source of truth for design decisions. If no specification exists, the plan is the authority.
|
|
106
|
+
|
|
107
|
+
**Important:** If research or discussion documents exist from earlier workflow phases, ignore them during implementation. They may contain outdated ideas, rejected approaches, or superseded decisions. The specification filtered and validated that content - refer only to the specification and plan.
|
|
96
108
|
|
|
97
109
|
## Project-Specific Conventions
|
|
98
110
|
|
|
@@ -9,23 +9,25 @@ Act as **expert technical architect**, **product owner**, and **plan documenter*
|
|
|
9
9
|
|
|
10
10
|
Your role spans product (WHAT we're building and WHY) and technical (HOW to structure the work).
|
|
11
11
|
|
|
12
|
-
##
|
|
12
|
+
## Purpose in the Workflow
|
|
13
13
|
|
|
14
|
-
|
|
15
|
-
|
|
16
|
-
|
|
17
|
-
4. **Planning** (YOU): HOW - phases, tasks, acceptance criteria
|
|
18
|
-
5. **Implementation** (next): DOING - tests first, then code
|
|
19
|
-
6. **Review** (final): VALIDATING - check work against artifacts
|
|
14
|
+
This skill can be used:
|
|
15
|
+
- **Sequentially** (Phase 4): From a validated specification
|
|
16
|
+
- **Standalone** (Contract entry): From any specification meeting format requirements
|
|
20
17
|
|
|
21
|
-
|
|
18
|
+
Either way: Transform specifications into actionable phases, tasks, and acceptance criteria.
|
|
22
19
|
|
|
23
|
-
|
|
20
|
+
### What This Skill Needs
|
|
21
|
+
|
|
22
|
+
- **Specification content** (required) - The validated decisions and requirements to plan from
|
|
23
|
+
- **Topic name** (optional) - Will derive from specification if not provided
|
|
24
|
+
- **Output format preference** (optional) - Will ask if not specified
|
|
24
25
|
|
|
25
|
-
|
|
26
|
-
- **Specification** (`docs/workflow/specification/{topic}.md`)
|
|
26
|
+
**If missing:** Will ask user for specification location or content.
|
|
27
27
|
|
|
28
|
-
|
|
28
|
+
## Source Material
|
|
29
|
+
|
|
30
|
+
**The specification is your sole input.** Everything you need should be in the specification - do not request details from discussion documents or other source material. If information is missing, ask for clarification on the specification itself.
|
|
29
31
|
|
|
30
32
|
## The Process
|
|
31
33
|
|
|
@@ -41,6 +43,8 @@ The specification is the **sole source of truth**. It contains validated, approv
|
|
|
41
43
|
|
|
42
44
|
**Commit frequently**: Commit at natural breaks, after significant exchanges, and before any context refresh. Context refresh = lost work.
|
|
43
45
|
|
|
44
|
-
**Never invent reasoning**: If it's not in the specification, ask again.
|
|
46
|
+
**Never invent reasoning**: If it's not in the specification, ask again. The specification is the golden document - all plan content must trace back to it.
|
|
45
47
|
|
|
46
48
|
**Create plans, not code**: Your job is phases, tasks, and acceptance criteria - not implementation.
|
|
49
|
+
|
|
50
|
+
**Collaborate with the user**: Planning is iterative. Stop and ask when the specification is ambiguous, multiple valid approaches exist, or you're uncertain about task scope. The user expects collaboration - don't guess when you can ask.
|
|
@@ -10,6 +10,20 @@ You are creating the formal implementation plan from the specification.
|
|
|
10
10
|
|
|
11
11
|
**Confirm output format with user.** Ask which format they want, then load the appropriate output adapter from the main skill file. If you don't know which format, ask.
|
|
12
12
|
|
|
13
|
+
## Planning is Collaborative
|
|
14
|
+
|
|
15
|
+
Planning is an iterative process between you and the user. The specification contains validated, considered decisions - planning translates that work into actionable structure. But translation still requires judgment, and the user should guide that judgment.
|
|
16
|
+
|
|
17
|
+
**Stop, reflect, and ask** when:
|
|
18
|
+
- The specification is ambiguous about implementation approach
|
|
19
|
+
- Multiple valid ways to structure phases or tasks exist
|
|
20
|
+
- You're uncertain whether a task is appropriately scoped
|
|
21
|
+
- Edge cases aren't fully addressed in the specification
|
|
22
|
+
|
|
23
|
+
**Never invent to fill gaps.** If the specification doesn't address something, use `[needs-info]` and ask the user. The specification is the golden document - everything in the plan must trace back to it.
|
|
24
|
+
|
|
25
|
+
**The user expects collaboration.** Planning doesn't have to be done by the agent alone. It's normal and encouraged to pause for guidance rather than guess.
|
|
26
|
+
|
|
13
27
|
## The Planning Process
|
|
14
28
|
|
|
15
29
|
### 1. Read Specification
|
|
@@ -86,13 +100,91 @@ Verify:
|
|
|
86
100
|
|
|
87
101
|
## Task Design
|
|
88
102
|
|
|
89
|
-
**Each task should**:
|
|
90
|
-
- Be a single TDD cycle
|
|
91
|
-
- Have micro acceptance (specific test name)
|
|
92
|
-
- Do one clear thing
|
|
93
|
-
|
|
94
103
|
**One task = One TDD cycle**: write test → implement → pass → commit
|
|
95
104
|
|
|
105
|
+
### Task Structure
|
|
106
|
+
|
|
107
|
+
Every task should follow this structure:
|
|
108
|
+
|
|
109
|
+
```markdown
|
|
110
|
+
### Task N: [Clear action statement]
|
|
111
|
+
|
|
112
|
+
**Problem**: Why this task exists - what issue or gap it addresses.
|
|
113
|
+
|
|
114
|
+
**Solution**: What we're building - the high-level approach.
|
|
115
|
+
|
|
116
|
+
**Outcome**: What success looks like - the verifiable end state.
|
|
117
|
+
|
|
118
|
+
**Do**:
|
|
119
|
+
- Specific implementation steps
|
|
120
|
+
- File locations and method names where helpful
|
|
121
|
+
- Concrete guidance, not vague directions
|
|
122
|
+
|
|
123
|
+
**Acceptance Criteria**:
|
|
124
|
+
- [ ] First verifiable criterion
|
|
125
|
+
- [ ] Second verifiable criterion
|
|
126
|
+
- [ ] Edge case handling criterion
|
|
127
|
+
|
|
128
|
+
**Tests**:
|
|
129
|
+
- `"it does the primary expected behavior"`
|
|
130
|
+
- `"it handles edge case correctly"`
|
|
131
|
+
- `"it fails appropriately for invalid input"`
|
|
132
|
+
|
|
133
|
+
**Context**: (when relevant)
|
|
134
|
+
> Relevant details from specification: code examples, architectural decisions,
|
|
135
|
+
> data models, or constraints that inform implementation.
|
|
136
|
+
```
|
|
137
|
+
|
|
138
|
+
### Field Requirements
|
|
139
|
+
|
|
140
|
+
| Field | Required | Notes |
|
|
141
|
+
|-------|----------|-------|
|
|
142
|
+
| Problem | Yes | One sentence minimum - why this task exists |
|
|
143
|
+
| Solution | Yes | One sentence minimum - what we're building |
|
|
144
|
+
| Outcome | Yes | One sentence minimum - what success looks like |
|
|
145
|
+
| Do | Yes | At least one concrete action |
|
|
146
|
+
| Acceptance Criteria | Yes | At least one pass/fail criterion |
|
|
147
|
+
| Tests | Yes | At least one test name; include edge cases, not just happy path |
|
|
148
|
+
| Context | When relevant | Only include when spec has details worth pulling forward |
|
|
149
|
+
|
|
150
|
+
### The Template as Quality Gate
|
|
151
|
+
|
|
152
|
+
If you struggle to articulate a clear Problem for a task, this signals the task may be:
|
|
153
|
+
|
|
154
|
+
- **Too granular**: Merge with a related task
|
|
155
|
+
- **Mechanical housekeeping**: Include as a step within another task
|
|
156
|
+
- **Poorly understood**: Revisit the specification
|
|
157
|
+
|
|
158
|
+
Every standalone task should have a reason to exist that can be stated simply. The template enforces this - difficulty completing it is diagnostic information, not a problem to work around.
|
|
159
|
+
|
|
160
|
+
### Vertical Slicing
|
|
161
|
+
|
|
162
|
+
Prefer **vertical slices** that deliver complete, testable functionality over horizontal slices that separate by technical layer.
|
|
163
|
+
|
|
164
|
+
**Horizontal (avoid)**:
|
|
165
|
+
```
|
|
166
|
+
Task 1: Create all database models
|
|
167
|
+
Task 2: Create all service classes
|
|
168
|
+
Task 3: Wire up integrations
|
|
169
|
+
Task 4: Add error handling
|
|
170
|
+
```
|
|
171
|
+
|
|
172
|
+
Nothing works until Task 4. No task is independently verifiable.
|
|
173
|
+
|
|
174
|
+
**Vertical (prefer)**:
|
|
175
|
+
```
|
|
176
|
+
Task 1: Fetch and store events from provider (happy path)
|
|
177
|
+
Task 2: Handle pagination for large result sets
|
|
178
|
+
Task 3: Handle authentication token refresh
|
|
179
|
+
Task 4: Handle rate limiting
|
|
180
|
+
```
|
|
181
|
+
|
|
182
|
+
Each task delivers a complete slice of functionality that can be tested in isolation.
|
|
183
|
+
|
|
184
|
+
Within a bounded feature, vertical slicing means each task completes a coherent unit of that feature's functionality - not that it must touch UI/API/database layers. The test is: *can this task be verified independently?*
|
|
185
|
+
|
|
186
|
+
TDD naturally encourages vertical slicing - when you think "what test can I write?", you frame work as complete, verifiable behavior rather than technical layers
|
|
187
|
+
|
|
96
188
|
## Plan as Source of Truth
|
|
97
189
|
|
|
98
190
|
The plan IS the source of truth. Every phase, every task must contain all information needed to execute it.
|
|
@@ -7,16 +7,20 @@ description: "Explore ideas, validate concepts, and research broadly across tech
|
|
|
7
7
|
|
|
8
8
|
Act as **research partner** with broad expertise spanning technical, product, business, and market domains. Your role is learning, exploration, and discovery.
|
|
9
9
|
|
|
10
|
-
##
|
|
10
|
+
## Purpose in the Workflow
|
|
11
11
|
|
|
12
|
-
|
|
13
|
-
|
|
14
|
-
|
|
15
|
-
4. **Planning** (after): HOW - phases, tasks, acceptance criteria
|
|
16
|
-
5. **Implementation** (after): DOING - tests first, then code
|
|
17
|
-
6. **Review** (final): VALIDATING - check work against artifacts
|
|
12
|
+
This skill can be used:
|
|
13
|
+
- **Sequentially** (Phase 1): First phase, to explore ideas before discussion
|
|
14
|
+
- **Standalone** (Contract entry): To research and validate any idea, feature, or concept
|
|
18
15
|
|
|
19
|
-
|
|
16
|
+
Either way: Explore feasibility (technical, business, market), validate assumptions, document findings.
|
|
17
|
+
|
|
18
|
+
### What This Skill Needs
|
|
19
|
+
|
|
20
|
+
- **Topic or idea** (required) - What to research/explore
|
|
21
|
+
- **Existing context** (optional) - Any prior research or constraints
|
|
22
|
+
|
|
23
|
+
**If missing:** Will ask user what they want to explore.
|
|
20
24
|
|
|
21
25
|
## Your Expertise
|
|
22
26
|
|
|
@@ -58,7 +62,7 @@ Start with one file. Early research is messy - topics aren't clear, you're follo
|
|
|
58
62
|
|
|
59
63
|
**Let themes emerge**: Over multiple sessions, topics may become distinct. When they do, split into semantic files (`market-landscape.md`, `technical-feasibility.md`).
|
|
60
64
|
|
|
61
|
-
**Periodic review**: Every few sessions, assess: are themes emerging? Split them out. Still fuzzy? Keep exploring. Ready for deeper discussion?
|
|
65
|
+
**Periodic review**: Every few sessions, assess: are themes emerging? Split them out. Still fuzzy? Keep exploring. Ready for deeper discussion? Research phase is complete.
|
|
62
66
|
|
|
63
67
|
## Documentation Loop
|
|
64
68
|
|
|
@@ -9,20 +9,37 @@ Act as a **senior software architect** with deep experience in code review. You
|
|
|
9
9
|
|
|
10
10
|
This is **product review**, **feature review**, **test review**, AND **code review**. Not just "does the code work?" but "was every task done correctly, tested properly, and built to professional standards?"
|
|
11
11
|
|
|
12
|
-
##
|
|
12
|
+
## Review Artifacts
|
|
13
13
|
|
|
14
|
-
|
|
15
|
-
|
|
16
|
-
3. **Specification** (artifact): REFINE - validated, standalone specification
|
|
17
|
-
4. **Planning** (artifact): HOW - phases, tasks, acceptance criteria
|
|
18
|
-
5. **Implementation** (completed): DOING - tests and code
|
|
19
|
-
6. **Review** (YOU): VALIDATING - verify every task
|
|
14
|
+
This skill reviews against available artifacts. Required:
|
|
15
|
+
- **Plan** (the tasks and acceptance criteria)
|
|
20
16
|
|
|
21
|
-
|
|
17
|
+
Optional but helpful:
|
|
18
|
+
- **Specification** (context for design decisions)
|
|
19
|
+
|
|
20
|
+
## Purpose in the Workflow
|
|
21
|
+
|
|
22
|
+
This skill can be used:
|
|
23
|
+
- **Sequentially** (Phase 6): After implementation of a planned feature
|
|
24
|
+
- **Standalone** (Contract entry): To review any implementation against a plan
|
|
25
|
+
|
|
26
|
+
Either way: Verify every plan task was implemented, tested adequately, and meets quality standards.
|
|
27
|
+
|
|
28
|
+
### What This Skill Needs
|
|
29
|
+
|
|
30
|
+
- **Plan content** (required) - Tasks and acceptance criteria to verify against
|
|
31
|
+
- **Specification content** (optional) - Context for design decisions
|
|
32
|
+
- **Implementation scope** (optional) - What code/files to review. Will identify from git if not specified.
|
|
33
|
+
|
|
34
|
+
**If missing:** Will ask user for plan location. Can proceed without specification.
|
|
22
35
|
|
|
23
36
|
## Review Approach
|
|
24
37
|
|
|
25
|
-
Start from the **plan** - it contains the granular tasks and acceptance criteria.
|
|
38
|
+
Start from the **plan** - it contains the granular tasks and acceptance criteria.
|
|
39
|
+
|
|
40
|
+
Use the **specification** for context if available. If no specification exists, the plan is the source of truth for design decisions.
|
|
41
|
+
|
|
42
|
+
Verify **all** tasks, not a sample.
|
|
26
43
|
|
|
27
44
|
```
|
|
28
45
|
Plan (tasks + acceptance criteria)
|
|
@@ -6,8 +6,10 @@
|
|
|
6
6
|
|
|
7
7
|
## Before Starting
|
|
8
8
|
|
|
9
|
-
1. Read plan
|
|
10
|
-
|
|
9
|
+
1. Read plan (from the location provided)
|
|
10
|
+
- If not found at expected path, ask user where the plan is
|
|
11
|
+
2. Read specification if available and linked in plan
|
|
12
|
+
- Not required, but helpful for context if it exists
|
|
11
13
|
3. Identify what code/files were changed
|
|
12
14
|
4. Check for project-specific skills in `.claude/skills/`
|
|
13
15
|
|
|
@@ -9,16 +9,23 @@ Act as **expert technical architect** and **specification builder**. Collaborate
|
|
|
9
9
|
|
|
10
10
|
Your role is to synthesize reference material, present it for validation, and build a specification that formal planning can execute against.
|
|
11
11
|
|
|
12
|
-
##
|
|
12
|
+
## Purpose in the Workflow
|
|
13
13
|
|
|
14
|
-
|
|
15
|
-
|
|
16
|
-
|
|
17
|
-
4. **Planning** (next): HOW - phases, tasks, acceptance criteria
|
|
18
|
-
5. **Implementation** (after): DOING - tests first, then code
|
|
19
|
-
6. **Review** (final): VALIDATING - check work against artifacts
|
|
14
|
+
This skill can be used:
|
|
15
|
+
- **Sequentially** (Phase 3): After discussion documents exist
|
|
16
|
+
- **Standalone** (Contract entry): With reference material from any source (research docs, conversation transcripts, design documents, inline feature description)
|
|
20
17
|
|
|
21
|
-
|
|
18
|
+
Either way: Transform unvalidated reference material into a specification that's **standalone and approved**.
|
|
19
|
+
|
|
20
|
+
### What This Skill Needs
|
|
21
|
+
|
|
22
|
+
- **Source material** (required) - The content to synthesize into a specification. Can be:
|
|
23
|
+
- Discussion document content (from sequential workflow)
|
|
24
|
+
- Inline feature description (from `/start-feature`)
|
|
25
|
+
- Any other reference material (requirements docs, transcripts, etc.)
|
|
26
|
+
- **Topic name** (required) - Used for the output filename
|
|
27
|
+
|
|
28
|
+
**If missing:** Will ask user to provide context or point to source files.
|
|
22
29
|
|
|
23
30
|
## The Process
|
|
24
31
|
|
|
@@ -26,7 +33,7 @@ You're at step 3. Build the specification. Don't jump to phases, tasks, or code.
|
|
|
26
33
|
|
|
27
34
|
**Output**: `docs/workflow/specification/{topic}.md`
|
|
28
35
|
|
|
29
|
-
**When complete**: User signs off
|
|
36
|
+
**When complete**: User signs off on the specification.
|
|
30
37
|
|
|
31
38
|
## What You Do
|
|
32
39
|
|