@sylphx/flow 2.18.2 → 2.18.3

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/CHANGELOG.md CHANGED
@@ -1,5 +1,11 @@
1
1
  # @sylphx/flow
2
2
 
3
+ ## 2.18.3 (2025-12-18)
4
+
5
+ ### ♻️ Refactoring
6
+
7
+ - **commands:** /continue as product thinking, not code checklist ([4f9cb0f](https://github.com/SylphxAI/flow/commit/4f9cb0f8eb3e0afca67b005bc94362d3582c6395))
8
+
3
9
  ## 2.18.2 (2025-12-18)
4
10
 
5
11
  ### ♻️ Refactoring
@@ -1,49 +1,62 @@
1
1
  ---
2
2
  name: continue
3
- description: Continue incomplete work - find gaps, finish features, fix what's broken
3
+ description: Iterate toward production-grade - find gaps from every angle, fix completely
4
4
  agent: coder
5
5
  ---
6
6
 
7
7
  # Continue
8
8
 
9
- Find what's incomplete. Finish it.
9
+ **Goal: Production-grade, not MVP.** Perfect architecture. Perfect naming. No workarounds.
10
10
 
11
- ## Mandate
11
+ ## Think From Every Perspective
12
12
 
13
- - **Think, don't checklist.** Understand the project. What would "done" look like?
14
- - **Delegate workers** for parallel research. You synthesize and verify.
15
- - **Fix, don't report.** Implement solutions directly.
13
+ Simulate being each persona. What's missing? What's frustrating? What breaks?
16
14
 
17
- ## Execution
15
+ - **New user** — First impression, onboarding, can they succeed without docs?
16
+ - **Power user** — Edge cases, advanced flows, what's limiting them?
17
+ - **Developer** — Is the code maintainable? Clear naming? Good abstractions?
18
+ - **Admin** — Can they manage users, debug issues, see what's happening?
19
+ - **Attacker** — Where are the security holes? What can be exploited?
20
+ - **3am oncall** — If it breaks, can they diagnose it? Are there logs? Alerts?
18
21
 
19
- 1. **Understand** Read README, entry points, core flows
22
+ Don't checklist. Actually **inhabit** each perspective and feel the friction.
20
23
 
21
- 2. **Find gaps** Not just TODO/FIXME, but:
22
- - Implicit broken (empty catch, happy path only, no logs)
23
- - Missing entirely (referenced but not implemented)
24
+ ## Delegate for Collective Wisdom
24
25
 
25
- 3. **Invoke skills** Before fixing, load guidelines for relevant domains:
26
- ```
27
- auth, account-security, billing, security, database, performance, observability...
28
- ```
29
- Skills contain: tech stack decisions, non-negotiables, patterns, anti-patterns.
26
+ Spawn multiple agents in parallel to ultrathink from different angles:
27
+ - One agent: user experience gaps
28
+ - One agent: security & trust gaps
29
+ - One agent: architecture & code quality gaps
30
+ - One agent: operability & observability gaps
30
31
 
31
- 4. **Fix by impact** What blocks users? Causes data loss? Fix those first. Fix completely.
32
+ You are the **Final Gate**. Synthesize their findings. Verify. Decide. Execute.
32
33
 
33
- 5. **Loop** — New gaps from fixes? → `/continue` again
34
+ ## Invoke Skills
35
+
36
+ Before fixing, load guidelines for relevant domains. Skills contain tech stack decisions, non-negotiables, patterns.
37
+
38
+ ## Fix Completely
39
+
40
+ No workarounds. No "good enough". Each fix should be the **final implementation** — production-ready, battle-tested quality.
41
+
42
+ ## Loop
43
+
44
+ After fixing: Are there new gaps? Did fixing X reveal Y was also broken?
45
+
46
+ If yes → `/continue` again. Keep iterating until production-grade.
34
47
 
35
48
  ## Output
36
49
 
37
50
  ```
38
- ## Found
39
- [What's incomplete and why]
51
+ ## Perspectives Explored
52
+ [Which personas, what gaps found from each]
40
53
 
41
54
  ## Fixed
42
- [Changes made]
55
+ [Changes made and why they're production-ready]
43
56
 
44
57
  ## Remains
45
- [Needs human decision]
58
+ [Needs human decision or blocked]
46
59
 
47
60
  ## Next
48
- [/continue | done]
61
+ [/continue | production-ready]
49
62
  ```
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "@sylphx/flow",
3
- "version": "2.18.2",
3
+ "version": "2.18.3",
4
4
  "description": "One CLI to rule them all. Unified orchestration layer for Claude Code, OpenCode, Cursor and all AI development tools. Auto-detection, auto-installation, auto-upgrade.",
5
5
  "type": "module",
6
6
  "bin": {