@tamsensedev/runner 0.13.0 → 0.18.0

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 (82) hide show
  1. package/LICENSE +5 -5
  2. package/README.md +61 -57
  3. package/dist/cli.js +646 -315
  4. package/dist/cli.js.map +1 -1
  5. package/dist/recorder.cjs.map +1 -1
  6. package/library/.claude-plugin/plugin.json +5 -5
  7. package/library/THIRD_PARTY_NOTICES.md +123 -123
  8. package/library/agents/code-architect.md +67 -67
  9. package/library/agents/code-explorer.md +61 -61
  10. package/library/agents/code-reviewer.md +73 -73
  11. package/library/agents/code-simplifier.md +62 -62
  12. package/library/agents/design-critic.md +52 -52
  13. package/library/agents/second-opinion.md +64 -64
  14. package/library/agents/security-auditor.md +86 -86
  15. package/library/agents/silent-failure-hunter.md +70 -70
  16. package/library/agents/test-coverage-analyzer.md +78 -78
  17. package/library/agents/test-runner.md +48 -48
  18. package/library/agents/ui-tester.md +50 -50
  19. package/library/licenses/Apache-2.0.txt +202 -202
  20. package/library/skills/accessibility/SKILL.md +144 -144
  21. package/library/skills/accessibility/references/wcag-checks.md +168 -168
  22. package/library/skills/addressing-review-feedback/SKILL.md +109 -109
  23. package/library/skills/app-screens/SKILL.md +96 -96
  24. package/library/skills/app-screens/references/shell-kpis-charts.md +108 -108
  25. package/library/skills/app-screens/references/states-density-phone.md +54 -54
  26. package/library/skills/app-screens/references/tables-forms-settings.md +104 -104
  27. package/library/skills/avoiding-generic-ai-design/SKILL.md +92 -92
  28. package/library/skills/avoiding-generic-ai-design/references/tells.md +112 -112
  29. package/library/skills/clean-commits/SKILL.md +85 -85
  30. package/library/skills/design-review/SKILL.md +107 -107
  31. package/library/skills/design-review/references/critique-passes.md +101 -101
  32. package/library/skills/design-review/references/screenshot-procedure.md +121 -121
  33. package/library/skills/differential-security-review/LICENSE +427 -427
  34. package/library/skills/differential-security-review/SKILL.md +105 -105
  35. package/library/skills/differential-security-review/references/adversarial.md +115 -115
  36. package/library/skills/differential-security-review/references/methodology.md +145 -145
  37. package/library/skills/differential-security-review/references/patterns.md +108 -108
  38. package/library/skills/false-positive-check/LICENSE +427 -427
  39. package/library/skills/false-positive-check/SKILL.md +95 -95
  40. package/library/skills/false-positive-check/references/bug-class-verification.md +128 -128
  41. package/library/skills/false-positive-check/references/deep-verification.md +113 -113
  42. package/library/skills/false-positive-check/references/evidence-templates.md +93 -93
  43. package/library/skills/false-positive-check/references/false-positive-patterns.md +122 -122
  44. package/library/skills/false-positive-check/references/gate-reviews.md +36 -36
  45. package/library/skills/false-positive-check/references/standard-verification.md +80 -80
  46. package/library/skills/focused-implementation/SKILL.md +104 -104
  47. package/library/skills/implementation-plan/SKILL.md +66 -66
  48. package/library/skills/interface-design/SKILL.md +144 -144
  49. package/library/skills/interface-design/references/craft-defaults.md +130 -130
  50. package/library/skills/interface-design/references/new-direction.md +62 -62
  51. package/library/skills/interface-design/references/writing.md +52 -52
  52. package/library/skills/interface-polish/SKILL.md +103 -103
  53. package/library/skills/interface-polish/references/detail-rules.md +154 -154
  54. package/library/skills/interface-polish/references/worst-case-data.md +158 -158
  55. package/library/skills/motion-design/SKILL.md +106 -106
  56. package/library/skills/motion-design/references/recipes.md +243 -243
  57. package/library/skills/postgres-performance/SKILL.md +97 -97
  58. package/library/skills/postgres-performance/references/rules.md +265 -265
  59. package/library/skills/project-orientation/SKILL.md +77 -77
  60. package/library/skills/reviewing-changes/SKILL.md +183 -183
  61. package/library/skills/safe-database-changes/SKILL.md +111 -111
  62. package/library/skills/safe-database-changes/references/backfills.md +75 -75
  63. package/library/skills/safe-database-changes/references/migration-tools.md +69 -69
  64. package/library/skills/safe-database-changes/references/postgres-ddl.md +199 -199
  65. package/library/skills/security-review/SKILL.md +132 -132
  66. package/library/skills/security-review/references/checks-by-area.md +107 -107
  67. package/library/skills/security-review/references/do-not-report.md +86 -86
  68. package/library/skills/spec-writing/SKILL.md +144 -144
  69. package/library/skills/systematic-debugging/SKILL.md +130 -130
  70. package/library/skills/systematic-debugging/references/flaky-tests.md +51 -51
  71. package/library/skills/systematic-debugging/references/root-cause-tracing.md +55 -55
  72. package/library/skills/task-interview/SKILL.md +84 -84
  73. package/library/skills/test-driven-development/SKILL.md +116 -116
  74. package/library/skills/test-driven-development/references/writing-good-tests.md +85 -85
  75. package/library/skills/ui-testing/SKILL.md +102 -102
  76. package/library/skills/ui-testing/references/headless-check.md +295 -295
  77. package/library/skills/ui-testing/references/local-stand.md +51 -51
  78. package/library/skills/ui-testing/references/playwright-patterns.md +134 -134
  79. package/library/skills/ui-testing/references/recording.md +61 -61
  80. package/library/skills/verification-before-completion/SKILL.md +113 -113
  81. package/library/skills/web-interface-guidelines/SKILL.md +202 -202
  82. package/package.json +46 -46
package/LICENSE CHANGED
@@ -1,5 +1,5 @@
1
- Copyright (c) 2026 TAMsense. All rights reserved.
2
-
3
- This package is provided to TAMsense customers to connect their computers to TAMsense. No license is granted to copy, modify or redistribute it outside that use.
4
-
5
- The skills and agents in the `library` folder adapt open source work under their own licenses. See `library/THIRD_PARTY_NOTICES.md`.
1
+ Copyright (c) 2026 TAMsense. All rights reserved.
2
+
3
+ This package is provided to TAMsense customers to connect their computers to TAMsense. No license is granted to copy, modify or redistribute it outside that use.
4
+
5
+ The skills and agents in the `library` folder adapt open source work under their own licenses. See `library/THIRD_PARTY_NOTICES.md`.
package/README.md CHANGED
@@ -1,57 +1,61 @@
1
- # @tamsensedev/runner
2
-
3
- Connects your computer to [TAMsense](https://my.tamsense.com), so the steps of your tasks run in Claude Code on your own machine, on your own Claude subscription.
4
-
5
- ## Requirements
6
-
7
- - Node.js 22 or newer
8
- - Claude Code, signed in (`claude` works in your terminal)
9
- - Git
10
-
11
- ## Connect this computer
12
-
13
- ```
14
- npx @tamsensedev/runner@latest login
15
- ```
16
-
17
- It shows a code and opens TAMsense in your browser. Confirm the code there, and the computer appears in your account under Computers.
18
-
19
- ## Work on your tasks
20
-
21
- Open a terminal in your code folder, then start the program there:
22
-
23
- ```
24
- cd <your code folder>
25
- npx @tamsensedev/runner@latest start
26
- ```
27
-
28
- It looks for code where you started it. If that folder is a git repository, it uses it. If not, it looks one level down, so starting in `C:\code` with `platform` and `wms` inside finds both. It does not look deeper, skips hidden folders and `node_modules`, and does not follow links. Each folder it finds is listed in the terminal (`Found code: platform (C:\code\platform)`) and on the site for this computer. A board whose repositories match a found folder uses it on this computer; a board without code offers it with one click. To add more, start the program from another code folder, or add a path in Board settings > Code.
29
-
30
- The computer takes the steps you own on the task board (questions and specs, code, review, tests) and works on them until you press Ctrl+C. You see everything Claude does live in the task chat, and you can message it or stop it there.
31
-
32
- When a task waits for something slow, such as a local stand that takes half an hour to start, the agent can set up a schedule: a short check on a timer, run by a cheap model on this computer for free. Each check writes one line in the task chat. While the computer sleeps or the program is stopped, its checks are skipped, not saved for later.
33
-
34
- ## Other commands
35
-
36
- ```
37
- npx @tamsensedev/runner@latest status show the connection
38
- npx @tamsensedev/runner@latest logout disconnect this computer
39
- npx @tamsensedev/runner@latest browser install the headless browser for interface checks
40
- npx @tamsensedev/runner@latest browser remove delete that browser
41
- ```
42
-
43
- Options:
44
-
45
- - `--name <name>` how this computer is called in TAMsense (login)
46
- - `--server <url>` TAMsense address, default https://my.tamsense.com (login)
47
- - `--parallel <n>` how many tasks to work on at once (start)
48
- - `--no-browser` do not install the headless browser for interface checks (start)
49
-
50
- ## What it does on your computer
51
-
52
- - Clones the task's repositories into `~/.tamsense/repos` and works in `~/.tamsense/work`.
53
- - Installs a headless browser for interface checks into `~/.tamsense` when `start` finds none, about 150 MB, without asking. If that fails, it says why and tasks still run without it.
54
- - Gets a short-lived GitHub token from TAMsense for those repositories only, and pushes the task branch. Pull requests are opened by TAMsense.
55
- - When a board takes its code from a folder on your computer (found at start, or added in board settings, Code tab), it reads only the git details of that folder to check it. Steps work in a git worktree of it in `~/.tamsense/work` on branch `task/<id>`: your working folder and its current branch stay as they are. If the folder has a remote, it fetches and pushes the task branch with your own git login; it never pushes the main branch. Without a remote the branch stays in the folder and you merge it yourself.
56
- - Records the browser Claude drives during a task (a replay of each page, and a video when asked) into `~/.tamsense/recordings`, attaches it to the task and deletes the local copy when the step ends. Requests from the tested site to the TAMsense tracker are blocked during these runs, so they never reach your analytics, and so are requests that change data (anything but GET, HEAD and OPTIONS) to addresses outside your computer.
57
- - Never stops processes it did not start.
1
+ # @tamsensedev/runner
2
+
3
+ Connects your computer to [TAMsense](https://my.tamsense.com), so the steps of your tasks run in Claude Code on your own machine, on your own Claude subscription.
4
+
5
+ ## Requirements
6
+
7
+ - Node.js 22 or newer
8
+ - Claude Code, signed in (`claude` works in your terminal)
9
+ - Git
10
+
11
+ ## Connect this computer
12
+
13
+ ```
14
+ npx @tamsensedev/runner@latest login
15
+ ```
16
+
17
+ It shows a code and opens TAMsense in your browser. Confirm the code there, and the computer appears in your account under Computers.
18
+
19
+ ## Work on your tasks
20
+
21
+ Open a terminal in your code folder, then start the program there:
22
+
23
+ ```
24
+ cd <your code folder>
25
+ npx @tamsensedev/runner@latest start
26
+ ```
27
+
28
+ It looks for code where you started it. If that folder is a git repository, it uses it. If not, it looks one level down, so starting in `C:\code` with `platform` and `wms` inside finds both. It does not look deeper, skips hidden folders and `node_modules`, and does not follow links. Each folder it finds is listed in the terminal (`Found code: platform (C:\code\platform)`) and on the site for this computer. A found folder joins the project by itself when one of its agents chose that repository, or when another computer in the project already has it; otherwise the site offers it with one click under this computer in Settings > Account. To add more, start the program from another code folder, or add a path there.
29
+
30
+ The computer takes the steps you own on the task board (questions and specs, code, review, tests) and works on them until you press Ctrl+C. You see everything Claude does live in the task chat, and you can message it or stop it there.
31
+
32
+ When a task waits for something slow, such as a local stand that takes half an hour to start, the agent can set up a schedule: a short check on a timer, run by a cheap model on this computer for free. Each check writes one line in the task chat. While the computer sleeps or the program is stopped, its checks are skipped, not saved for later.
33
+
34
+ When you write to Sense in a task chat, Sense decides on this computer with the Claude signed in here, for free, next to the tasks it works on. It gets the task as you see it on the site and no files or commands. Questions about analytics or visitor sessions it asks TAMsense cloud, which answers in plain words and charges your AI credits for that answer. The terminal shows one line per answer, `Sense answered on task #12`, never your words. If this computer does not answer in time, the cloud answers instead.
35
+
36
+ Routines can run here too: a job that repeats on a schedule, such as finding stuck orders every morning. A routine run makes tasks and asks people on TAMsense. When its routine may fix things, every action that changes something goes through the safety rules of the project and waits for a person to allow it, just like a task step. A test run changes nothing and asks nobody.
37
+
38
+ ## Other commands
39
+
40
+ ```
41
+ npx @tamsensedev/runner@latest status show the connection
42
+ npx @tamsensedev/runner@latest logout disconnect this computer
43
+ npx @tamsensedev/runner@latest browser install the headless browser for interface checks
44
+ npx @tamsensedev/runner@latest browser remove delete that browser
45
+ ```
46
+
47
+ Options:
48
+
49
+ - `--name <name>` how this computer is called in TAMsense (login)
50
+ - `--server <url>` TAMsense address, default https://my.tamsense.com (login)
51
+ - `--parallel <n>` how many tasks to work on at once (start)
52
+ - `--no-browser` do not install the headless browser for interface checks (start)
53
+
54
+ ## What it does on your computer
55
+
56
+ - Clones the task's repositories into `~/.tamsense/repos` and works in `~/.tamsense/work`.
57
+ - Installs a headless browser for interface checks into `~/.tamsense` when `start` finds none, about 150 MB, without asking. If that fails, it says why and tasks still run without it.
58
+ - Gets a short-lived GitHub token from TAMsense for those repositories only, and pushes the task branch. Pull requests are opened by TAMsense.
59
+ - When an agent takes its code from a folder on your computer (found at start, or added in Settings > Account under this computer), it reads only the git details of that folder to check it. Steps work in a git worktree of it in `~/.tamsense/work` on branch `task/<id>`: your working folder and its current branch stay as they are. If the folder has a remote, it fetches and pushes the task branch with your own git login; it never pushes the main branch. Without a remote the branch stays in the folder and you merge it yourself.
60
+ - Records the browser Claude drives during a task (a replay of each page, and a video when asked) into `~/.tamsense/recordings`, attaches it to the task and deletes the local copy when the step ends. Requests from the tested site to the TAMsense tracker are blocked during these runs, so they never reach your analytics, and so are requests that change data (anything but GET, HEAD and OPTIONS) to addresses outside your computer.
61
+ - Never stops processes it did not start.