@veolab/discoverylab 1.4.0 → 1.4.1

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.
@@ -12,7 +12,7 @@
12
12
  "name": "discoverylab",
13
13
  "source": ".",
14
14
  "description": "AI-powered app testing & marketing asset generator. Record mobile/web apps, run automated tests with Maestro & Playwright, and generate professional screenshots, GIFs, and test reports.",
15
- "version": "1.4.0",
15
+ "version": "1.4.1",
16
16
  "author": {
17
17
  "name": "Anderson Melo"
18
18
  },
@@ -1,7 +1,7 @@
1
1
  {
2
2
  "name": "discoverylab",
3
3
  "description": "AI-powered app testing & marketing asset generator. Record mobile/web apps, run automated tests with Maestro & Playwright, and generate professional screenshots, GIFs, and test reports.",
4
- "version": "1.4.0",
4
+ "version": "1.4.1",
5
5
  "author": {
6
6
  "name": "Anderson Melo",
7
7
  "email": "anderson.90@gmail.com"
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "@veolab/discoverylab",
3
- "version": "1.4.0",
3
+ "version": "1.4.1",
4
4
  "description": "AI-powered app testing & evidence generator - Claude Code Plugin",
5
5
  "type": "module",
6
6
  "main": "dist/index.js",
@@ -68,7 +68,7 @@
68
68
  "claude-plugin": {
69
69
  "name": "DiscoveryLab",
70
70
  "description": "AI-powered app testing & evidence generator",
71
- "version": "1.4.0",
71
+ "version": "1.4.1",
72
72
  "tools": [
73
73
  "dlab.capture.screen",
74
74
  "dlab.capture.emulator",
@@ -15,50 +15,67 @@ tags:
15
15
 
16
16
  # DiscoveryLab Knowledge Brain
17
17
 
18
- When the user asks about app flows, screens, UI elements, or how something works in an app they've captured, use DiscoveryLab's knowledge base to find answers.
18
+ DiscoveryLab captures app recordings and analyzes them with OCR + AI, building a knowledge base of every flow, screen, and UI element. This skill lets you query that knowledge.
19
19
 
20
20
  ## When to Use
21
21
 
22
- Use `dlab.knowledge.search` when the user:
23
- - Asks about a specific app screen ("how does the login work?")
24
- - Asks about UI elements ("what buttons are on the paywall?")
25
- - Asks about user flows ("what's the onboarding flow?")
26
- - References an app by name ("in StyliApp, what does the closet screen show?")
27
- - Needs context about captured recordings
28
- - Compares flows or screens
22
+ Use this whenever the user asks about:
23
+ - How a specific screen or flow works in their app
24
+ - What UI elements exist on a page (buttons, inputs, labels)
25
+ - The user journey through a feature (onboarding, checkout, settings)
26
+ - Comparing different captures of the same flow
27
+ - Any reference to app screens, recordings, or captured content
28
+ - Context about what was tested or recorded
29
+
30
+ Also use proactively when:
31
+ - The user is working on code related to a feature that was captured
32
+ - You need visual context about the app to give better answers
33
+ - The user mentions a Jira ticket that might be linked to a project
29
34
 
30
35
  ## Tools
31
36
 
32
37
  ### `dlab.knowledge.search`
33
- Search across all captured projects. Matches against project names, AI analysis, OCR text, tags, and linked tickets.
38
+ Semantic search across all captured projects. Matches against: project names, AI analysis summaries, OCR text from every screen, tags, and linked tickets.
34
39
 
35
40
  ```
36
- dlab.knowledge.search { query: "login StyliApp" }
37
- dlab.knowledge.search { query: "paywall premium" }
41
+ dlab.knowledge.search { query: "login" }
42
+ dlab.knowledge.search { query: "paywall" }
38
43
  dlab.knowledge.search { query: "onboarding flow" }
44
+ dlab.knowledge.search { query: "PROJ-123" }
45
+ dlab.knowledge.search { query: "settings profile" }
39
46
  ```
40
47
 
41
- Returns: project overview, user flow steps, UI elements found, OCR text sample, and project ID for deeper lookup.
48
+ Returns per match:
49
+ - App overview (what the screen/flow does)
50
+ - User flow steps (numbered sequence)
51
+ - UI elements found (buttons, inputs, navigation)
52
+ - OCR text sample (actual text visible on screens)
53
+ - Project ID for deeper lookup
42
54
 
43
55
  ### `dlab.knowledge.summary`
44
- Get a high-level overview of all captured app knowledge.
56
+ High-level overview of the entire knowledge base - all projects grouped by app with stats.
45
57
 
46
58
  ```
47
59
  dlab.knowledge.summary {}
48
60
  ```
49
61
 
50
- Returns: all projects grouped by app, with stats (flows count, screens, analysis status).
62
+ Use this when:
63
+ - The user asks "what do we have captured?"
64
+ - You need to orient yourself on what apps/flows exist
65
+ - Starting a new conversation and need context
51
66
 
52
67
  ### `dlab.project.get`
53
- For full details on a specific project found via search.
68
+ Full details on a specific project found via search. Use when the search result summary is not enough.
54
69
 
55
70
  ```
56
- dlab.project.get { projectId: "abc123" }
71
+ dlab.project.get { projectId: "<id from search results>" }
57
72
  ```
58
73
 
59
- ## Behavior
74
+ ## How to Respond
60
75
 
61
- 1. Always search first before saying you don't know about an app flow
62
- 2. If no match, show available projects so the user knows what's captured
63
- 3. Use `dlab.project.get` for full details when the search result needs more depth
64
- 4. Multiple projects may cover the same flow (different versions) - present the most recent first
76
+ 1. **Search first** - never say "I don't have information about that" without searching
77
+ 2. **No match?** - run `dlab.knowledge.summary` to show the user what's available
78
+ 3. **Multiple results** - present the most recent first, note if there are different versions
79
+ 4. **Cite the source** - mention which project/recording the information comes from
80
+ 5. **Suggest captures** - if the user asks about a flow that doesn't exist, suggest they capture it with DiscoveryLab
81
+ 6. **Be specific** - use the OCR text and UI elements from results to give precise answers, not generic ones