create-sdd-project 0.9.5 → 0.9.6

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.
Files changed (3) hide show
  1. package/package.json +1 -1
  2. package/template/.claude/commands/review-plan.md +58 -22
  3. package/template/.gemini/commands/review-plan-instructions.md +54 -17
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "create-sdd-project",
3
- "version": "0.9.5",
3
+ "version": "0.9.6",
4
4
  "description": "Create a new SDD DevFlow project with AI-assisted development workflow",
5
5
  "bin": {
6
6
  "create-sdd-project": "bin/cli.js"
package/template/.claude/commands/review-plan.md CHANGED
@@ -1,52 +1,88 @@
1
- Review the Implementation Plan in the current ticket using an external model for independent critique.
1
+ Review the Implementation Plan in the current ticket using external models for independent critique.
2
2
 
3
3
  ## Prerequisites
4
4
 
5
- - An external AI CLI tool installed: [Codex CLI](https://github.com/openai/codex), [Gemini CLI](https://github.com/google-gemini/gemini-cli), or similar
6
5
  - An active feature with a completed Implementation Plan (Step 2)
6
+ - Ideally, one or more external AI CLIs installed: [Gemini CLI](https://github.com/google-gemini/gemini-cli), [Codex CLI](https://github.com/openai/codex), or similar
7
7
 
8
8
  ## What to do
9
9
 
10
10
  1. **Find the current ticket** — Read `docs/project_notes/product-tracker.md` → Active Session → ticket path
11
- 2. **Extract the plan** — Read the ticket's `## Implementation Plan` section
12
- 3. **Send to external reviewer** — Run the external CLI in read-only mode with this prompt:
13
11
 
12
+ 2. **Detect available reviewers** — Check which external CLIs are installed:
13
+
14
+ ```bash
15
+ which gemini 2>/dev/null && echo "gemini: available" || echo "gemini: not found"
16
+ which codex 2>/dev/null && echo "codex: available" || echo "codex: not found"
14
17
  ```
15
- You are reviewing an Implementation Plan for a software feature. Your job is to find problems, not praise.
16
18
 
17
- Review the plan below and report:
18
- 1. **Errors** — Wrong assumptions, impossible steps, missing dependencies
19
- 2. **Gaps** — Missing error handling, edge cases, rollback scenarios
20
- 3. **Vagueness** — Steps too ambiguous to implement with TDD (no clear input/output)
21
- 4. **Over-engineering** — Unnecessary abstractions, premature optimization
22
- 5. **Order issues** — Steps that depend on later steps
19
+ 3. **Prepare the review input** — Extract the spec and plan into a temp file with the review prompt. Use the feature ID from the Active Session (e.g., `F023`):
20
+
21
+ ```bash
22
+ TICKET=$(ls docs/tickets/F023-*.md) # Use the feature ID from Step 1
23
+
24
+ cat > /tmp/review-prompt.txt <<'CRITERIA'
25
+ You are reviewing an Implementation Plan for a software feature. Your job is to find real problems, not praise. But if the plan is solid, say APPROVED — do not manufacture issues that are not there.
26
+
27
+ Below you will find the Spec (what to build) and the Implementation Plan (how to build it). Review the plan and report:
28
+ 1. Errors — Wrong assumptions, impossible steps, missing dependencies, plan contradicts the spec
29
+ 2. Gaps — Missing error handling, edge cases, rollback scenarios
30
+ 3. Vagueness — Steps too ambiguous to implement with TDD (no clear input/output)
31
+ 4. Over-engineering — Unnecessary abstractions, premature optimization
32
+ 5. Order issues — Steps that depend on later steps
23
33
 
24
34
  For each issue, state: [CRITICAL/IMPORTANT/SUGGESTION] — description — proposed fix.
25
35
 
26
36
  End with: VERDICT: APPROVED | VERDICT: REVISE (if any CRITICAL or 2+ IMPORTANT issues)
27
37
 
28
38
  ---
29
- PLAN:
30
- <paste the Implementation Plan here>
39
+ SPEC AND PLAN:
40
+ CRITERIA
41
+
42
+ sed -n '/## Spec/,/## Acceptance Criteria/p' "$TICKET" >> /tmp/review-prompt.txt
31
43
  ```
32
44
 
33
- Example with Codex CLI:
45
+ 4. **Send for review** — Execute **only one** of the following paths based on Step 2 results:
46
+
47
+ ### Path A: Both CLIs available (best — two independent perspectives)
48
+
34
49
  ```bash
35
- codex exec -m o4-mini --quiet "Review this Implementation Plan: $(cat docs/tickets/FXXX-*.md | sed -n '/## Implementation Plan/,/## Acceptance Criteria/p')"
50
+ cat /tmp/review-prompt.txt | gemini > /tmp/review-gemini.txt 2>&1 &
51
+ cat /tmp/review-prompt.txt | codex exec - > /tmp/review-codex.txt 2>&1 &
52
+ wait
53
+
54
+ echo "=== GEMINI REVIEW ===" && cat /tmp/review-gemini.txt
55
+ echo "=== CODEX REVIEW ===" && cat /tmp/review-codex.txt
36
56
  ```
37
57
 
38
- Example with Gemini CLI:
58
+ Consolidate findings — issues flagged by both models independently carry higher weight. Deduplicate and prioritize.
59
+
60
+ ### Path B: One CLI available
61
+
39
62
  ```bash
40
- gemini -m gemini-2.5-pro "Review this Implementation Plan: $(cat docs/tickets/FXXX-*.md | sed -n '/## Implementation Plan/,/## Acceptance Criteria/p')"
63
+ # Gemini only
64
+ cat /tmp/review-prompt.txt | gemini
65
+
66
+ # Codex only
67
+ cat /tmp/review-prompt.txt | codex exec -
41
68
  ```
42
69
 
43
- 4. **Process the review** — If VERDICT: REVISE, update the plan addressing CRITICAL and IMPORTANT issues
44
- 5. **Optional second round** — Send the revised plan for a final audit if significant changes were made
45
- 6. **Log the review** — Add a note in the ticket's Completion Log: "Plan reviewed by [model] — N issues found, N addressed"
70
+ ### Path C: No external CLI available (self-review fallback)
71
+
72
+ If no external CLI is installed, perform the review yourself. Re-read the full Implementation Plan from the ticket, then review it with this mindset:
73
+
74
+ > Review the plan one more time trying to find any flaw, anything we might have forgotten, or any improvement that is missing. This is a very important part of the project and it is worth doing well. Do your best effort.
75
+
76
+ Apply the same 5 criteria from the prompt above. For each issue, state severity, description, and proposed fix. End with VERDICT.
77
+
78
+ 5. **Process the review** — If any VERDICT is REVISE, update the plan addressing CRITICAL and IMPORTANT issues
79
+ 6. **Optional second round** — Send the revised plan for a final audit if significant changes were made
80
+ 7. **Log the review** — Add a note in the ticket's Completion Log: "Plan reviewed by [model(s) or self-review] — N issues found, N addressed"
46
81
 
47
82
  ## Notes
48
83
 
49
84
  - This command is **optional** — the workflow's built-in Plan Self-Review (Step 2.4) always runs automatically
50
85
  - Most valuable for **Complex** features or plans with 8+ steps
51
- - The external model runs read-only — it reviews text, not code
52
- - If no external CLI is available, you can copy-paste the plan into a different AI chat for manual review
86
+ - External models run read-only — they review text, not code
87
+ - Both CLIs use their latest default model when no `-m` flag is specified — no need to hardcode model names
88
+ - Path C (self-review) is a last resort — external review gives genuinely independent perspectives that self-review cannot
package/template/.gemini/commands/review-plan-instructions.md CHANGED
@@ -1,23 +1,33 @@
1
1
  ## Review Plan — Instructions
2
2
 
3
- Review the Implementation Plan in the current ticket using an external model for independent critique.
3
+ Review the Implementation Plan in the current ticket using external models for independent critique.
4
4
 
5
5
  ### Prerequisites
6
6
 
7
- - An external AI CLI tool installed: Codex CLI, Claude Code, or similar
8
7
  - An active feature with a completed Implementation Plan (Step 2)
8
+ - Ideally, one or more external AI CLIs installed: Codex CLI, Claude Code, or similar
9
9
 
10
10
  ### Steps
11
11
 
12
12
  1. **Find the current ticket** — Read `docs/project_notes/product-tracker.md` → Active Session → ticket path
13
- 2. **Extract the plan** — Read the ticket's `## Implementation Plan` section
14
- 3. **Send to external reviewer** — Run the external CLI in read-only mode with this prompt:
15
13
 
14
+ 2. **Detect available reviewers** — Check which external CLIs are installed:
15
+
16
+ ```bash
17
+ which claude 2>/dev/null && echo "claude: available" || echo "claude: not found"
18
+ which codex 2>/dev/null && echo "codex: available" || echo "codex: not found"
16
19
  ```
17
- You are reviewing an Implementation Plan for a software feature. Your job is to find problems, not praise.
18
20
 
19
- Review the plan below and report:
20
- 1. Errors — Wrong assumptions, impossible steps, missing dependencies
21
+ 3. **Prepare the review input** — Extract the spec and plan into a temp file with the review prompt. Use the feature ID from the Active Session (e.g., `F023`):
22
+
23
+ ```bash
24
+ TICKET=$(ls docs/tickets/F023-*.md) # Use the feature ID from Step 1
25
+
26
+ cat > /tmp/review-prompt.txt <<'CRITERIA'
27
+ You are reviewing an Implementation Plan for a software feature. Your job is to find real problems, not praise. But if the plan is solid, say APPROVED — do not manufacture issues that are not there.
28
+
29
+ Below you will find the Spec (what to build) and the Implementation Plan (how to build it). Review the plan and report:
30
+ 1. Errors — Wrong assumptions, impossible steps, missing dependencies, plan contradicts the spec
21
31
  2. Gaps — Missing error handling, edge cases, rollback scenarios
22
32
  3. Vagueness — Steps too ambiguous to implement with TDD (no clear input/output)
23
33
  4. Over-engineering — Unnecessary abstractions, premature optimization
@@ -28,26 +38,53 @@ For each issue, state: [CRITICAL/IMPORTANT/SUGGESTION] — description — propo
28
38
  End with: VERDICT: APPROVED | VERDICT: REVISE (if any CRITICAL or 2+ IMPORTANT issues)
29
39
 
30
40
  ---
31
- PLAN:
32
- <paste the Implementation Plan here>
41
+ SPEC AND PLAN:
42
+ CRITERIA
43
+
44
+ sed -n '/## Spec/,/## Acceptance Criteria/p' "$TICKET" >> /tmp/review-prompt.txt
33
45
  ```
34
46
 
35
- Example with Claude Code:
47
+ 4. **Send for review** — Execute **only one** of the following paths based on Step 2 results:
48
+
49
+ #### Path A: Both CLIs available (best — two independent perspectives)
50
+
36
51
  ```bash
37
- claude -m claude-sonnet-4-20250514 --print "Review this plan: $(cat docs/tickets/FXXX-*.md | sed -n '/## Implementation Plan/,/## Acceptance Criteria/p')"
52
+ cat /tmp/review-prompt.txt | claude --print > /tmp/review-claude.txt 2>&1 &
53
+ cat /tmp/review-prompt.txt | codex exec - > /tmp/review-codex.txt 2>&1 &
54
+ wait
55
+
56
+ echo "=== CLAUDE REVIEW ===" && cat /tmp/review-claude.txt
57
+ echo "=== CODEX REVIEW ===" && cat /tmp/review-codex.txt
38
58
  ```
39
59
 
40
- Example with Codex CLI:
60
+ Consolidate findings — issues flagged by both models independently carry higher weight. Deduplicate and prioritize.
61
+
62
+ #### Path B: One CLI available
63
+
41
64
  ```bash
42
- codex exec -m o4-mini --quiet "Review this plan: $(cat docs/tickets/FXXX-*.md | sed -n '/## Implementation Plan/,/## Acceptance Criteria/p')"
65
+ # Claude only
66
+ cat /tmp/review-prompt.txt | claude --print
67
+
68
+ # Codex only
69
+ cat /tmp/review-prompt.txt | codex exec -
43
70
  ```
44
71
 
45
- 4. **Process the review** — If VERDICT: REVISE, update the plan addressing CRITICAL and IMPORTANT issues
46
- 5. **Optional second round** — Send the revised plan for a final audit if significant changes were made
47
- 6. **Log the review** — Add a note in the ticket's Completion Log: "Plan reviewed by [model] — N issues found, N addressed"
72
+ #### Path C: No external CLI available (self-review fallback)
73
+
74
+ If no external CLI is installed, perform the review yourself. Re-read the full Implementation Plan from the ticket, then review it with this mindset:
75
+
76
+ > Review the plan one more time trying to find any flaw, anything we might have forgotten, or any improvement that is missing. This is a very important part of the project and it is worth doing well. Do your best effort.
77
+
78
+ Apply the same 5 criteria from the prompt above. For each issue, state severity, description, and proposed fix. End with VERDICT.
79
+
80
+ 5. **Process the review** — If any VERDICT is REVISE, update the plan addressing CRITICAL and IMPORTANT issues
81
+ 6. **Optional second round** — Send the revised plan for a final audit if significant changes were made
82
+ 7. **Log the review** — Add a note in the ticket's Completion Log: "Plan reviewed by [model(s) or self-review] — N issues found, N addressed"
48
83
 
49
84
  ### Notes
50
85
 
51
86
  - This command is **optional** — the workflow's built-in Plan Self-Review (Step 2.4) always runs automatically
52
87
  - Most valuable for Complex features or plans with 8+ steps
53
- - The external model runs read-only — it reviews text, not code
88
+ - External models run read-only — they review text, not code
89
+ - Both CLIs use their latest default model when no `-m` flag is specified — no need to hardcode model names
90
+ - Path C (self-review) is a last resort — external review gives genuinely independent perspectives that self-review cannot