crewforth 3.0.0 → 3.0.1-rc.2

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 (44) hide show
  1. package/CHANGELOG.md +112 -0
  2. package/README.md +1 -1
  3. package/VERSION +1 -1
  4. package/adopt.sh +222 -19
  5. package/kit/CLAUDE.md +5 -4
  6. package/kit/README.md +1 -1
  7. package/kit/agents/crew-frontend-expert.md +6 -6
  8. package/kit/agents/crew-session-manager.md +1 -1
  9. package/kit/eval/doctor.sh +77 -18
  10. package/kit/eval/golden-routing.txt +15 -0
  11. package/kit/eval/lib/git-bash.sh +108 -0
  12. package/kit/eval/preflight.sh +2 -2
  13. package/kit/eval/smoke-test.sh +761 -29
  14. package/kit/hooks/board-sync.sh +1 -1
  15. package/kit/hooks/board.sh +4 -3
  16. package/kit/hooks/guard-bash.sh +195 -21
  17. package/kit/hooks/guard-commit-scan.sh +2 -2
  18. package/kit/hooks/guard-powershell.sh +328 -0
  19. package/kit/hooks/guard-write.sh +28 -3
  20. package/kit/hooks/route-hint.sh +5 -1
  21. package/kit/hooks/session-update-check.sh +2 -2
  22. package/kit/hooks/skill-trust.sh +57 -14
  23. package/kit/legacy-blobs.tsv +208 -0
  24. package/kit/settings.json +21 -0
  25. package/kit/skills/automode-policy/SKILL.md +1 -1
  26. package/kit/skills/automode-policy/scripts/check.sh +2 -2
  27. package/kit/skills/crew-board/SKILL.md +1 -1
  28. package/kit/skills/crew-doctor/SKILL.md +13 -5
  29. package/kit/skills/crew-gates/SKILL.md +2 -1
  30. package/kit/skills/crew-handoff/SKILL.md +1 -1
  31. package/kit/skills/crew-skill/SKILL.md +2 -2
  32. package/kit/skills/crew-studio/SKILL.md +3 -2
  33. package/kit/skills/crew-update/SKILL.md +3 -3
  34. package/kit/skills/frontend/SKILL.md +26 -2
  35. package/kit/skills/frontend-rn-expo/SKILL.md +3 -0
  36. package/kit/skills/handoff/SKILL.md +2 -2
  37. package/kit/skills/reflect/SKILL.md +2 -2
  38. package/kit/skills/teamboard/SKILL.md +1 -1
  39. package/kit/skills/token-budget/SKILL.md +2 -2
  40. package/kit/skills/trace-scan/SKILL.md +4 -3
  41. package/kit/studio/README.md +1 -1
  42. package/kit/studio/server/lib/permissions.js +3 -0
  43. package/package.json +5 -4
  44. package/start.sh +15 -8
package/kit/legacy-blobs.tsv ADDED
@@ -0,0 +1,208 @@
1
+ # Generated by packaging/gen-legacy-blobs.sh from git tags v1.0.0..v2.13.0 — do not edit by hand.
2
+ # skills/cqrs-aop-module and skills/devarch-module rows decide trust only; the legacy sweep never moves them.
3
+ # component install_path blob_sha1 first_tag last_tag source
4
+ agents/backend-expert-cck.md .claude/agents/backend-expert-cck.md c22b19c5927a3c178ac9f00034bdc0c584cbcb9e v1.0.0 v1.0.4 shipped
5
+ agents/backend-expert-cck.md .claude/agents/backend-expert-cck.md e57a61131b9be4aa7c6af740eb856b8b2e94c70f v1.0.0 v1.0.4 generic-variant
6
+ agents/backend-expert-csk.md .claude/agents/backend-expert-csk.md 00f1919476923e6a418c6d7d33d60d4c8a11197c v2.12.0 v2.12.0 shipped
7
+ agents/backend-expert-csk.md .claude/agents/backend-expert-csk.md 0491fc1fc3613a6e117f1d8cb64f7b34db57a05b v2.13.0 v2.13.0 shipped
8
+ agents/backend-expert-csk.md .claude/agents/backend-expert-csk.md 1d0f9e6714b448f5fa35a31e4871739474ffb9d2 v2.10.0 v2.10.1 generic-variant
9
+ agents/backend-expert-csk.md .claude/agents/backend-expert-csk.md 1d8cff0669911e32def6536eaa82e01de32941fd v2.11.0 v2.11.0 generic-variant
10
+ agents/backend-expert-csk.md .claude/agents/backend-expert-csk.md 47af8cc46c4498004d8e623998634f9c76e5d2cc v1.0.5 v1.0.9 shipped
11
+ agents/backend-expert-csk.md .claude/agents/backend-expert-csk.md 4f6d0d1fa7d9ed69c6ab9655570b08fb2bbb846b v1.8.0 v1.10.0 shipped
12
+ agents/backend-expert-csk.md .claude/agents/backend-expert-csk.md 784334e66cfbf419d6d3283717ee67768d613241 v1.0.5 v1.7.0 generic-variant
13
+ agents/backend-expert-csk.md .claude/agents/backend-expert-csk.md 7d852ef3e84021546697397a806a7d42b16cf6bb v1.4.0 v1.4.4 shipped
14
+ agents/backend-expert-csk.md .claude/agents/backend-expert-csk.md 8909d2d7f89e21dcccc459d71bbccbc791803773 v1.5.0 v1.7.0 shipped
15
+ agents/backend-expert-csk.md .claude/agents/backend-expert-csk.md 9b3feb7ba5ab66c09b943d1b8479cdbb65f429b8 v2.0.0 v2.1.0 shipped
16
+ agents/backend-expert-csk.md .claude/agents/backend-expert-csk.md a2f357d57b4dd56c5eb21fe08bd354ea94bcfc78 v2.2.0 v2.9.0 shipped
17
+ agents/backend-expert-csk.md .claude/agents/backend-expert-csk.md ae8185a746dd6fa3847654e35f80ad1b7c51aa79 v2.0.0 v2.9.0 generic-variant
18
+ agents/backend-expert-csk.md .claude/agents/backend-expert-csk.md b29658bbb6df7b8296add6e1a9e6d5b8f45cb627 v2.11.0 v2.11.0 shipped
19
+ agents/backend-expert-csk.md .claude/agents/backend-expert-csk.md b2e0b8d4350f17335e5b57afc5b52558c9123ab8 v1.1.0 v1.1.4 shipped
20
+ agents/backend-expert-csk.md .claude/agents/backend-expert-csk.md c4ad12e59a1bdc3a18c0d32e597ac1878d11762c v1.1.5 v1.3.0 shipped
21
+ agents/backend-expert-csk.md .claude/agents/backend-expert-csk.md cdf4063f57b0d6fb7a5d298bd5dad264a0ae8da9 v2.10.0 v2.10.1 shipped
22
+ agents/backend-expert-csk.md .claude/agents/backend-expert-csk.md cf4de23586c6a45f106575e2b476b8a753fa898d v1.10.1 v1.10.1 shipped
23
+ agents/backend-expert-csk.md .claude/agents/backend-expert-csk.md d6dbb69552be86241055b2008aebac46b208b4ea v1.8.0 v1.10.0 generic-variant
24
+ agents/backend-expert-csk.md .claude/agents/backend-expert-csk.md dcc481833dcd7a327e006876b1b32676f54b2b14 v1.10.1 v1.10.1 generic-variant
25
+ agents/backend-expert-csk.md .claude/agents/backend-expert-csk.md e0b95b4511b199567245011c2da2e5276d74b3e2 v2.12.0 v2.13.0 generic-variant
26
+ agents/commit-agent-cck.md .claude/agents/commit-agent-cck.md f18f42b2699ffdf322e353ccef9f94f8236051f0 v1.0.0 v1.0.4 shipped
27
+ agents/commit-agent-csk.md .claude/agents/commit-agent-csk.md 50ed57c5e622346b0c4b66f5a8c6ec09652948c3 v1.10.1 v1.10.1 shipped
28
+ agents/commit-agent-csk.md .claude/agents/commit-agent-csk.md 52d5ba72b09638961af979af78c7eea2e8cba4e4 v1.1.0 v1.3.0 shipped
29
+ agents/commit-agent-csk.md .claude/agents/commit-agent-csk.md 78679dad967ba46f063be8789b2197f80b0d685c v1.0.5 v1.0.9 shipped
30
+ agents/commit-agent-csk.md .claude/agents/commit-agent-csk.md 9440a68f1358bdd8b8abfc87fa69c8e0e6551c4a v2.12.0 v2.12.0 shipped
31
+ agents/commit-agent-csk.md .claude/agents/commit-agent-csk.md 98473fd52d92e955a453b0fcef43c2b9a86da27d v2.11.0 v2.11.0 shipped
32
+ agents/commit-agent-csk.md .claude/agents/commit-agent-csk.md ab4fde6ea4ec19768421c08f089dcef82345ee9b v2.0.0 v2.10.1 shipped
33
+ agents/commit-agent-csk.md .claude/agents/commit-agent-csk.md c6014a5aed1e2b1ccac6923efa316476feb6c263 v1.4.0 v1.10.0 shipped
34
+ agents/commit-agent-csk.md .claude/agents/commit-agent-csk.md dc7ba83d8d776113a875c93e42feb51e31dd844a v2.13.0 v2.13.0 shipped
35
+ agents/database-expert-cck.md .claude/agents/database-expert-cck.md 119e2f63a7a9b1ffe144871afeddc573fe8d16c4 v1.0.0 v1.0.4 shipped
36
+ agents/database-expert-csk.md .claude/agents/database-expert-csk.md 166eac7c2cc5b6306a902ec0ea0d846b4972f476 v1.10.1 v2.9.0 shipped
37
+ agents/database-expert-csk.md .claude/agents/database-expert-csk.md 3408d922d8590049c06ba0ff30e1806f0143e506 v2.12.0 v2.13.0 shipped
38
+ agents/database-expert-csk.md .claude/agents/database-expert-csk.md 46a057f4e9bd305ff2d26442a7ce0078dea13512 v1.0.5 v1.0.9 shipped
39
+ agents/database-expert-csk.md .claude/agents/database-expert-csk.md 5f637f8170eaf6476250b78fb59e19f6d27e7503 v1.1.0 v1.3.0 shipped
40
+ agents/database-expert-csk.md .claude/agents/database-expert-csk.md 65014f1666ace6ef305496cb09d2f0616fcf1954 v2.11.0 v2.11.0 shipped
41
+ agents/database-expert-csk.md .claude/agents/database-expert-csk.md 77d1116406a87cc181e38e9bfb58a7f756ae1bfe v1.4.0 v1.4.4 shipped
42
+ agents/database-expert-csk.md .claude/agents/database-expert-csk.md 80f0abfda9ea8cd78b737b3bf1e20e49fe6222e2 v2.10.0 v2.10.1 shipped
43
+ agents/database-expert-csk.md .claude/agents/database-expert-csk.md 99baa79a4198a8c2d8f000cc2ffceb64600999db v1.5.0 v1.7.0 shipped
44
+ agents/database-expert-csk.md .claude/agents/database-expert-csk.md f170af1c3fd92463334b4fd28d42243257cc466f v1.8.0 v1.10.0 shipped
45
+ agents/devops-expert-cck.md .claude/agents/devops-expert-cck.md b638568a16f84348a58f9ec220730acd2b7afdbc v1.0.0 v1.0.4 shipped
46
+ agents/devops-expert-csk.md .claude/agents/devops-expert-csk.md 18b093be9691c0f585730ee14c4aef309e3afd35 v2.7.0 v2.11.0 shipped
47
+ agents/devops-expert-csk.md .claude/agents/devops-expert-csk.md 394b4293196f9d483624df41f550e1ecbc83f197 v1.4.0 v1.4.4 shipped
48
+ agents/devops-expert-csk.md .claude/agents/devops-expert-csk.md 44f8dc6e53006a5a1369f42966e08cd148bdab07 v1.0.5 v1.0.9 shipped
49
+ agents/devops-expert-csk.md .claude/agents/devops-expert-csk.md 58731401392db22d59a14fd656a1f3e7a144a42d v2.12.0 v2.13.0 shipped
50
+ agents/devops-expert-csk.md .claude/agents/devops-expert-csk.md 66cd189adc53442640a4f6b13330d6703c1ff023 v1.1.0 v1.3.0 shipped
51
+ agents/devops-expert-csk.md .claude/agents/devops-expert-csk.md 86af53b1b7d3f4ddc4c258584060a3b494369b51 v1.10.1 v2.6.0 shipped
52
+ agents/devops-expert-csk.md .claude/agents/devops-expert-csk.md 9c0aa0729f6e5f43027059c8592b1c448c4490b7 v1.5.0 v1.7.0 shipped
53
+ agents/devops-expert-csk.md .claude/agents/devops-expert-csk.md f83ceabb6c05f15f7773c5b2ba674643546b60a2 v1.8.0 v1.10.0 shipped
54
+ agents/frontend-expert-cck.md .claude/agents/frontend-expert-cck.md 64e6c4a827fafa83868e4055ebeaa1fc01224c98 v1.0.0 v1.0.4 shipped
55
+ agents/frontend-expert-csk.md .claude/agents/frontend-expert-csk.md 014b26ef30afdfb04dcf0c77da1ab5773c46377b v2.10.0 v2.10.1 shipped
56
+ agents/frontend-expert-csk.md .claude/agents/frontend-expert-csk.md 2c00558d0704be36ee0785c0eb734066a514b820 v1.5.0 v1.7.0 shipped
57
+ agents/frontend-expert-csk.md .claude/agents/frontend-expert-csk.md 5be44682dae96e8d9a6abc3c20f6a733b248ddc9 v1.1.0 v1.3.0 shipped
58
+ agents/frontend-expert-csk.md .claude/agents/frontend-expert-csk.md 5e8942252feccaede36be9a36c5945155ea4ec86 v1.0.5 v1.0.9 shipped
59
+ agents/frontend-expert-csk.md .claude/agents/frontend-expert-csk.md 850f77803ac0e18b134193a42fc6559637c141b6 v1.8.0 v1.10.0 shipped
60
+ agents/frontend-expert-csk.md .claude/agents/frontend-expert-csk.md 9868fe2552aac73b933707b909d36949dc92a283 v1.10.1 v2.6.0 shipped
61
+ agents/frontend-expert-csk.md .claude/agents/frontend-expert-csk.md b723a93dc24d6b428b333169d593f0c524e3a76f v1.4.0 v1.4.4 shipped
62
+ agents/frontend-expert-csk.md .claude/agents/frontend-expert-csk.md cc59c3cf55cf370f9a5c2aec16cdedeea8641180 v2.7.0 v2.9.0 shipped
63
+ agents/frontend-expert-csk.md .claude/agents/frontend-expert-csk.md d28bd4e1fbc98ba1adfd089f5aa4dc5a7826d049 v2.12.0 v2.13.0 shipped
64
+ agents/frontend-expert-csk.md .claude/agents/frontend-expert-csk.md de3c12987545a68cbf9f5f41f0fc58bb28c730e5 v2.11.0 v2.11.0 shipped
65
+ agents/performance-expert-csk.md .claude/agents/performance-expert-csk.md 532900a6097ad59132eefb4a627a149dd25cfda2 v2.0.0 v2.3.1 shipped
66
+ agents/performance-expert-csk.md .claude/agents/performance-expert-csk.md 9650c02094e0e85a7fd27416ad6741e6b65b44d1 v1.10.1 v1.10.1 shipped
67
+ agents/performance-expert-csk.md .claude/agents/performance-expert-csk.md b312c848a63f6031f1be172afaa381d8351bb7bd v2.7.0 v2.11.0 shipped
68
+ agents/performance-expert-csk.md .claude/agents/performance-expert-csk.md cb127f372743fc444ee24d006ab53fb2e3c82a16 v2.12.0 v2.13.0 shipped
69
+ agents/performance-expert-csk.md .claude/agents/performance-expert-csk.md cf91da3ee7ac0c44928e3b19fd81b9d297b58d04 v1.8.0 v1.10.0 shipped
70
+ agents/performance-expert-csk.md .claude/agents/performance-expert-csk.md ddb716e0b0fffd8ea89210b27cbdff65ed14412f v2.3.2 v2.6.0 shipped
71
+ agents/planner-cck.md .claude/agents/planner-cck.md 8616eb7be63f95b6a2e9eb8c0479307f1ee44ab9 v1.0.0 v1.0.4 shipped
72
+ agents/planner-csk.md .claude/agents/planner-csk.md 074251a2877330bdd41083f67120277fc5f23c34 v1.2.0 v1.2.1 shipped
73
+ agents/planner-csk.md .claude/agents/planner-csk.md 2a82d44d6638e915d02dc291eb456a35a43c4a60 v1.4.0 v1.4.4 shipped
74
+ agents/planner-csk.md .claude/agents/planner-csk.md 4f59056e65caaa4af53fa8ebec29bb03a6c22159 v1.1.0 v1.1.12 shipped
75
+ agents/planner-csk.md .claude/agents/planner-csk.md 7ac0c1878153294b70867102eee4150014dfaa7f v2.0.0 v2.2.2 shipped
76
+ agents/planner-csk.md .claude/agents/planner-csk.md 8a45e96780bc9564d59a835d447adcc6da556222 v2.3.0 v2.13.0 shipped
77
+ agents/planner-csk.md .claude/agents/planner-csk.md 920de2c777a42e35f4e3c1f5176c4b66cba01722 v1.10.1 v1.10.1 shipped
78
+ agents/planner-csk.md .claude/agents/planner-csk.md d2854353bd898e87631a47d4a92b895704ce5f2f v1.2.2 v1.3.0 shipped
79
+ agents/planner-csk.md .claude/agents/planner-csk.md deb7b6d916f3fba7709f21841990116bdee19a43 v1.5.0 v1.10.0 shipped
80
+ agents/planner-csk.md .claude/agents/planner-csk.md ed99bbcda67c2a7d755e9c955f1bd292cd5c4871 v1.0.5 v1.0.9 shipped
81
+ agents/privacy-agent-cck.md .claude/agents/privacy-agent-cck.md ae8742c5f99c5dfffaaf953ab18bbeb83486bbb3 v1.0.0 v1.0.4 shipped
82
+ agents/privacy-agent-csk.md .claude/agents/privacy-agent-csk.md 0a93bfc5d226818353b7801eba7afb76b73e10d2 v1.5.0 v1.10.0 shipped
83
+ agents/privacy-agent-csk.md .claude/agents/privacy-agent-csk.md 36dc80a35ce9c54fc146893e8dfa3a78375e4d61 v1.0.5 v1.0.9 shipped
84
+ agents/privacy-agent-csk.md .claude/agents/privacy-agent-csk.md 4891580c9d889eba05dc2f93f282911b6a63ea2b v1.4.0 v1.4.4 shipped
85
+ agents/privacy-agent-csk.md .claude/agents/privacy-agent-csk.md 7c72c9e704b916fd2f5453dd0509d6cef4142830 v2.3.2 v2.13.0 shipped
86
+ agents/privacy-agent-csk.md .claude/agents/privacy-agent-csk.md 7eb6c03ec937d98cb6c25669efc54540162a95be v2.3.1 v2.3.1 shipped
87
+ agents/privacy-agent-csk.md .claude/agents/privacy-agent-csk.md 83847f138cbc594ec36c6751d2aabd897d012a77 v1.10.1 v1.10.1 shipped
88
+ agents/privacy-agent-csk.md .claude/agents/privacy-agent-csk.md d06d042d1a2244cdf9d9d7c161fb4a22c1e845dc v2.0.0 v2.3.0 shipped
89
+ agents/privacy-agent-csk.md .claude/agents/privacy-agent-csk.md eab4496b03652d794a3188b5279564c64def1bae v1.1.0 v1.3.0 shipped
90
+ agents/review-agent-cck.md .claude/agents/review-agent-cck.md dcecad08983e949d0f79bd98fb707c4347c36aa5 v1.0.0 v1.0.4 shipped
91
+ agents/review-agent-csk.md .claude/agents/review-agent-csk.md 0d756e596d46c106ef402cd4ff882a4c63a2a65f v2.12.0 v2.13.0 shipped
92
+ agents/review-agent-csk.md .claude/agents/review-agent-csk.md 166762f0dee9bbb4645deb9591cba6fad5f77d63 v1.4.0 v1.4.4 shipped
93
+ agents/review-agent-csk.md .claude/agents/review-agent-csk.md 1acdfe4abc06769ecfa175f24c18e2ea3531a5fb v1.0.5 v1.0.9 shipped
94
+ agents/review-agent-csk.md .claude/agents/review-agent-csk.md 67a045446a7fdcd4c34b726b30a49fe9ec331d32 v2.7.0 v2.9.0 shipped
95
+ agents/review-agent-csk.md .claude/agents/review-agent-csk.md 9569e7e2eca589c124145ba50afc8cdf68f405a8 v1.2.0 v1.3.0 shipped
96
+ agents/review-agent-csk.md .claude/agents/review-agent-csk.md 9da1a4e1ca7a2d9a818c1545ceb9e3a4bf7fbbcf v1.10.1 v1.10.1 shipped
97
+ agents/review-agent-csk.md .claude/agents/review-agent-csk.md a263d93577485dd6a5737d6c9148917b066c6cd0 v2.10.0 v2.11.0 shipped
98
+ agents/review-agent-csk.md .claude/agents/review-agent-csk.md bdc117db60e7f8fff198157b86a67e5a2a6f2fff v2.0.0 v2.3.1 shipped
99
+ agents/review-agent-csk.md .claude/agents/review-agent-csk.md c772dbaeb97cb9c7e5d778ae1a63e8b6b7ce218b v1.5.0 v1.10.0 shipped
100
+ agents/review-agent-csk.md .claude/agents/review-agent-csk.md ddc6e062893ac548f3f3e8bb7456436a1c109732 v1.1.0 v1.1.12 shipped
101
+ agents/review-agent-csk.md .claude/agents/review-agent-csk.md e8c03e70dcb11f4864d5a1fce91c217d169261d2 v2.3.2 v2.6.0 shipped
102
+ agents/security-expert-cck.md .claude/agents/security-expert-cck.md 68b89986b7093c82974e53e9f3e68ba21010e3a8 v1.0.0 v1.0.4 shipped
103
+ agents/security-expert-csk.md .claude/agents/security-expert-csk.md 0f13f6858a5851edbd06fa4d0355fcfa206180aa v1.0.5 v1.0.9 shipped
104
+ agents/security-expert-csk.md .claude/agents/security-expert-csk.md 2a6177409ec5f0683ebd66bcc8eef999e6b1bfb9 v1.5.0 v1.6.3 shipped
105
+ agents/security-expert-csk.md .claude/agents/security-expert-csk.md 3e1bd87b690232e645ef03ab816982dfbdba7d22 v1.7.0 v1.10.0 shipped
106
+ agents/security-expert-csk.md .claude/agents/security-expert-csk.md 5cd1898649ccc9369a72d0458437438e1641e5cf v1.10.1 v1.10.1 shipped
107
+ agents/security-expert-csk.md .claude/agents/security-expert-csk.md 9839f7af8ba0fc623777590314aeccedfe0284f9 v1.1.0 v1.3.0 shipped
108
+ agents/security-expert-csk.md .claude/agents/security-expert-csk.md adb263d88c660d7bda4d8a32865e7ee72bade850 v2.3.2 v2.11.0 shipped
109
+ agents/security-expert-csk.md .claude/agents/security-expert-csk.md adf670fa194e2d395e33b000c9837fdaeb51b64e v2.0.0 v2.3.1 shipped
110
+ agents/security-expert-csk.md .claude/agents/security-expert-csk.md d83f7192044318c71064d9c5a0468e9e6b44602e v2.12.0 v2.13.0 shipped
111
+ agents/security-expert-csk.md .claude/agents/security-expert-csk.md e85bd5f82410580dc14b178e8599453cc85ca49f v1.4.0 v1.4.4 shipped
112
+ agents/session-manager-cck.md .claude/agents/session-manager-cck.md fee3b91860ff7b8706f403585ec5f899029c7228 v1.0.0 v1.0.4 shipped
113
+ agents/session-manager-csk.md .claude/agents/session-manager-csk.md 022e305778ec474819d80fa9e745b998bbad82b6 v2.12.0 v2.13.0 shipped
114
+ agents/session-manager-csk.md .claude/agents/session-manager-csk.md 112df7120221d3c1e6ea699ae582bea09bb87d89 v1.0.5 v1.0.9 shipped
115
+ agents/session-manager-csk.md .claude/agents/session-manager-csk.md 229f4561a685eaf726ea3543eaffc184d2164360 v2.3.0 v2.10.1 shipped
116
+ agents/session-manager-csk.md .claude/agents/session-manager-csk.md 28caec646014d1def3f09006a7bf58eaa717d40d v2.0.0 v2.2.2 shipped
117
+ agents/session-manager-csk.md .claude/agents/session-manager-csk.md b48628f4b6ea4e8ecbe8e957538887ef4b70c8fe v1.4.0 v1.10.0 shipped
118
+ agents/session-manager-csk.md .claude/agents/session-manager-csk.md bf71533b52ca87a9bb4f0d59683e3990668597ac v2.11.0 v2.11.0 shipped
119
+ agents/session-manager-csk.md .claude/agents/session-manager-csk.md c1ecde0b4569e4ae0e003f79bcd9175559214e2a v1.10.1 v1.10.1 shipped
120
+ agents/session-manager-csk.md .claude/agents/session-manager-csk.md c37e5e76677fdc52fd666baa9bb15855b4134f04 v1.1.0 v1.3.0 shipped
121
+ agents/test-expert-cck.md .claude/agents/test-expert-cck.md 91a4c8d80e2d0c2bf8f3769614902a1b7e32deb4 v1.0.0 v1.0.4 shipped
122
+ agents/test-expert-csk.md .claude/agents/test-expert-csk.md 3c00e0e26f5750b2105a23011366244c8fee0980 v1.4.0 v1.4.4 shipped
123
+ agents/test-expert-csk.md .claude/agents/test-expert-csk.md 51d6309cb4cb94bcdf7528aaf0643ec9d7d74d1b v1.5.0 v1.7.0 shipped
124
+ agents/test-expert-csk.md .claude/agents/test-expert-csk.md 5557ab3799c76d6cd4a41bd35a1df1ca9d998d65 v2.10.0 v2.11.0 shipped
125
+ agents/test-expert-csk.md .claude/agents/test-expert-csk.md aa3a142846611f461dd93ac56b0777a0756da756 v1.10.1 v2.9.0 shipped
126
+ agents/test-expert-csk.md .claude/agents/test-expert-csk.md bc89e60e07d258498466a07609a45bc4e3e9bdd6 v1.8.0 v1.10.0 shipped
127
+ agents/test-expert-csk.md .claude/agents/test-expert-csk.md e431bd8ff2945eedcf47ae622a253b47936100ea v2.12.0 v2.13.0 shipped
128
+ agents/test-expert-csk.md .claude/agents/test-expert-csk.md e76b14bf7df3ebb7d6cb6701723b98642f5e2d55 v1.0.5 v1.0.9 shipped
129
+ agents/test-expert-csk.md .claude/agents/test-expert-csk.md f07f8d233403063897eeb1b2985d8a8a9ad5ab87 v1.1.0 v1.3.0 shipped
130
+ commands/board-csk.md .claude/commands/board-csk.md 1ecc7ca78ed7974c5c2cee44d84426972d0c1b8d v2.3.0 v2.13.0 shipped
131
+ commands/brainstorm-csk.md .claude/commands/brainstorm-csk.md 47e3c3eef01df809d03bcda6c49590e51aee68e0 v2.0.0 v2.13.0 shipped
132
+ commands/brainstorm.md .claude/commands/brainstorm.md 1bad6554355522f086e1d77684810961083306d8 v1.10.1 v1.10.1 shipped
133
+ commands/brainstorm.md .claude/commands/brainstorm.md c78f3e9af4d6a07c010f6fd21b118ddcb3e4733a v1.2.0 v1.10.0 shipped
134
+ commands/doctor-csk.md .claude/commands/doctor-csk.md 4ac401d2060714df0d4c8749fd89f0a360673922 v1.8.0 v2.6.0 shipped
135
+ commands/doctor-csk.md .claude/commands/doctor-csk.md 584b947df9d74b8b394b72e02cb871d803fc56bf v2.9.0 v2.10.0 shipped
136
+ commands/doctor-csk.md .claude/commands/doctor-csk.md 6382439c036be5018925c2cbeaa777acc001e866 v1.4.0 v1.7.0 shipped
137
+ commands/doctor-csk.md .claude/commands/doctor-csk.md 9ae81d6f002a3a31a33b44992fb15a7631e49524 v2.10.1 v2.13.0 shipped
138
+ commands/doctor-csk.md .claude/commands/doctor-csk.md bbf9d3dd7285f483a3c0dda2259f6bdda8192a50 v2.7.0 v2.8.0 shipped
139
+ commands/gates-csk.md .claude/commands/gates-csk.md 6075778c6b11ed6ba881180100f9ca89065cd876 v2.5.0 v2.13.0 shipped
140
+ commands/handoff-csk.md .claude/commands/handoff-csk.md 80d5725dec1c71f5396564a5012d791a017a7d7c v2.0.0 v2.6.0 shipped
141
+ commands/handoff-csk.md .claude/commands/handoff-csk.md c611fafa8f1bef5bb39df8b3115f3ac7ca4ae8c6 v2.7.0 v2.13.0 shipped
142
+ commands/handoff.md .claude/commands/handoff.md 4b49650835149ae5d8d232ffb719b2f6485499b3 v1.0.0 v1.10.1 shipped
143
+ commands/plan-csk.md .claude/commands/plan-csk.md 83647040a1f237fe7d7e85923aa6963b86a94176 v2.0.0 v2.13.0 shipped
144
+ commands/plan.md .claude/commands/plan.md 16cc2075d73240ab071bcdb5d879a0ef355fc459 v1.0.0 v1.0.4 shipped
145
+ commands/plan.md .claude/commands/plan.md 572e51a0bffecfdb695e103b16d9ff5473ed6328 v1.10.1 v1.10.1 shipped
146
+ commands/plan.md .claude/commands/plan.md 594598e0e115cbfba29b4cc8673e200b12401524 v1.0.5 v1.10.0 shipped
147
+ commands/review-csk.md .claude/commands/review-csk.md 786c6a92e1c31c52f267e8a73743f904bd86ef69 v2.0.0 v2.11.0 shipped
148
+ commands/review-csk.md .claude/commands/review-csk.md ee877f5ca53317bcd00366d64e512864f0a8b178 v2.12.0 v2.13.0 shipped
149
+ commands/review.md .claude/commands/review.md 2cc6b766c8ca6c0b2b753a32fc1756884b7e9929 v1.0.0 v1.0.4 shipped
150
+ commands/review.md .claude/commands/review.md 52854ea2208172660cf74d8780596f06ceb3c127 v1.8.0 v1.10.0 shipped
151
+ commands/review.md .claude/commands/review.md f1f7da311dab08383be7fe4454d1495afdc1a848 v1.0.5 v1.7.0 shipped
152
+ commands/review.md .claude/commands/review.md f39ecf9b8dcf355b47f6450a636fb81bddbabc78 v1.10.1 v1.10.1 shipped
153
+ commands/ship-csk.md .claude/commands/ship-csk.md b0b2a3ff4d1a7aa32f2affbdfce8d4e7346e9fae v2.3.0 v2.13.0 shipped
154
+ commands/ship-csk.md .claude/commands/ship-csk.md fccc42de47ea9c06beaf12396df978a85b870131 v2.0.0 v2.2.2 shipped
155
+ commands/ship.md .claude/commands/ship.md 0b107bb1b052c02a82a5ed1df797232458920b66 v1.10.1 v1.10.1 shipped
156
+ commands/ship.md .claude/commands/ship.md c603ad465e2709685e63837621e96b9a21a29036 v1.0.5 v1.10.0 shipped
157
+ commands/ship.md .claude/commands/ship.md ecde12be572a435925a95c1fefa6fd3997bb9300 v1.0.0 v1.0.4 shipped
158
+ commands/simplify.md .claude/commands/simplify.md 5a4e630b7736822e89149add9c639623339d6d52 v1.10.1 v1.10.1 shipped
159
+ commands/simplify.md .claude/commands/simplify.md 8885a6f24081068f377c3b2bdf97b21870e4fe17 v1.0.0 v1.10.0 shipped
160
+ commands/skill-csk.md .claude/commands/skill-csk.md 4b03d450dd56c62371b737006f9650d3fe98d02e v2.7.0 v2.11.0 shipped
161
+ commands/skill-csk.md .claude/commands/skill-csk.md aeab6d0d3f56f23748e5307387c76e60183763c3 v2.12.0 v2.13.0 shipped
162
+ commands/studio-csk.md .claude/commands/studio-csk.md 1e0a9fcbc6dbaa897fa378184ee9435ffa6a9c25 v2.11.0 v2.12.0 shipped
163
+ commands/studio-csk.md .claude/commands/studio-csk.md 24b19e8e39f78b5a2c88358c0c9dfdc36230c9ac v2.9.0 v2.10.0 shipped
164
+ commands/studio-csk.md .claude/commands/studio-csk.md 66d8ba78d2b5c9f07a8f8e48364f78c6cef3937b v2.13.0 v2.13.0 shipped
165
+ commands/studio-csk.md .claude/commands/studio-csk.md 7835907eb2bdc53d814c54c5eccb85847bf7bea2 v2.10.1 v2.10.1 shipped
166
+ commands/update-csk.md .claude/commands/update-csk.md 730f176930afc38c41bb172e9aa505352b44e675 v1.4.0 v1.4.1 shipped
167
+ commands/update-csk.md .claude/commands/update-csk.md 8cb138fa62ededb3eaa07f08cecca504f30d132f v1.4.2 v1.6.2 shipped
168
+ commands/update-csk.md .claude/commands/update-csk.md d90707474bde09c2d694ea01902e68d293e978d7 v1.6.3 v2.13.0 shipped
169
+ skills/code-review-csk .claude/skills/code-review-csk/SKILL.md 1d3568f3b7bd7b18a809dbedce8a5d2c961642a0 v2.3.2 v2.6.0 shipped
170
+ skills/code-review-csk .claude/skills/code-review-csk/SKILL.md 4a274bb80c43d0003f402382e91b336568ec6ec8 v2.12.0 v2.13.0 shipped
171
+ skills/code-review-csk .claude/skills/code-review-csk/SKILL.md 8f92d592a4c42ba25f1c0f9b83a65e176a9d783b v2.0.0 v2.3.1 shipped
172
+ skills/code-review-csk .claude/skills/code-review-csk/SKILL.md dfcbcb148bfe4c92a30f94058672b162e92710d5 v2.7.0 v2.11.0 shipped
173
+ skills/code-review-csk .claude/skills/code-review-csk/references/history-search.md 2bcf3b0f5f82bc9d9325b4916552f00661e26e47 v2.12.0 v2.13.0 shipped
174
+ skills/code-review-csk .claude/skills/code-review-csk/references/panel-mode.md 373c39f61aa7be1f95b8a8b0e85c9d8858703013 v2.0.0 v2.13.0 shipped
175
+ skills/code-review-csk .claude/skills/code-review-csk/references/receiving.md 5cd7d5e627c9ad165ec7317571b7e95fdd1e75e2 v2.12.0 v2.13.0 shipped
176
+ skills/code-review .claude/skills/code-review/SKILL.md 16621ab9fbed5eb86a012cd395e81309fcf50283 v1.2.1 v1.3.0 shipped
177
+ skills/code-review .claude/skills/code-review/SKILL.md 804fa3c93ddbc6eee09200c82f4fd8c3a60e3ce2 v1.0.5 v1.0.9 shipped
178
+ skills/code-review .claude/skills/code-review/SKILL.md 8bcc6dcbac1857b605a7f62674a9329c986a73e6 v1.4.0 v1.10.1 shipped
179
+ skills/code-review .claude/skills/code-review/SKILL.md 8f05626bd45bbd9067ceeb219e5e0f3c52cbf9dd v1.1.0 v1.1.12 shipped
180
+ skills/code-review .claude/skills/code-review/SKILL.md 92a99014c3393845042a90367a38d83055d9fc06 v1.0.0 v1.0.4 shipped
181
+ skills/code-review .claude/skills/code-review/SKILL.md db18b839c5ba4b2b3c68c0edee84ed339fb08548 v1.2.0 v1.2.0 shipped
182
+ skills/code-review .claude/skills/code-review/references/panel-mode.md 373c39f61aa7be1f95b8a8b0e85c9d8858703013 v1.4.0 v1.10.1 shipped
183
+ skills/cqrs-aop-module .claude/skills/cqrs-aop-module/SKILL.md 1ee0c303296f17e192ed6708fb5c05ab77177777 v2.0.0 v2.10.1 devarch-renamed
184
+ skills/cqrs-aop-module .claude/skills/cqrs-aop-module/SKILL.md 3799cff8565213300c2dd6c0ec73d4c1a342dcc7 v1.0.5 v1.0.9 devarch-renamed
185
+ skills/cqrs-aop-module .claude/skills/cqrs-aop-module/SKILL.md 3cfb7e2065c383675e684d0b4c34bfb8248b15bb v2.11.0 v2.11.0 devarch-renamed
186
+ skills/cqrs-aop-module .claude/skills/cqrs-aop-module/SKILL.md 47624b783a54a32fe3a6adece8c8aeff5942194f v1.0.0 v1.0.4 devarch-renamed
187
+ skills/cqrs-aop-module .claude/skills/cqrs-aop-module/SKILL.md 48cfd2d06b9cd76a268bf21544044d09bcb904d5 v2.12.0 v2.13.0 shipped
188
+ skills/cqrs-aop-module .claude/skills/cqrs-aop-module/SKILL.md 699cc57b63c84e7c57a50e7fc025b63a31303822 v1.4.0 v1.10.1 devarch-renamed
189
+ skills/cqrs-aop-module .claude/skills/cqrs-aop-module/SKILL.md aae3d007e856bf4a5b4aeaeb74ecf5577b0b5789 v1.1.0 v1.1.4 devarch-renamed
190
+ skills/cqrs-aop-module .claude/skills/cqrs-aop-module/SKILL.md fdfafb0d0127daec51a04612be1513d3bded343f v1.1.5 v1.3.0 devarch-renamed
191
+ skills/cqrs-aop-module .claude/skills/cqrs-aop-module/references/skeleton.md fe73b6431ef362d7e7610fe2b14c3a5e4baf8d37 v1.4.0 v2.13.0 shipped+devarch-renamed
192
+ skills/devarch-module .claude/skills/devarch-module/SKILL.md 1ee0c303296f17e192ed6708fb5c05ab77177777 v2.0.0 v2.10.1 shipped
193
+ skills/devarch-module .claude/skills/devarch-module/SKILL.md 3799cff8565213300c2dd6c0ec73d4c1a342dcc7 v1.0.5 v1.0.9 shipped
194
+ skills/devarch-module .claude/skills/devarch-module/SKILL.md 3cfb7e2065c383675e684d0b4c34bfb8248b15bb v2.11.0 v2.11.0 shipped
195
+ skills/devarch-module .claude/skills/devarch-module/SKILL.md 47624b783a54a32fe3a6adece8c8aeff5942194f v1.0.0 v1.0.4 shipped
196
+ skills/devarch-module .claude/skills/devarch-module/SKILL.md 699cc57b63c84e7c57a50e7fc025b63a31303822 v1.4.0 v1.10.1 shipped
197
+ skills/devarch-module .claude/skills/devarch-module/SKILL.md aae3d007e856bf4a5b4aeaeb74ecf5577b0b5789 v1.1.0 v1.1.4 shipped
198
+ skills/devarch-module .claude/skills/devarch-module/SKILL.md fdfafb0d0127daec51a04612be1513d3bded343f v1.1.5 v1.3.0 shipped
199
+ skills/devarch-module .claude/skills/devarch-module/references/skeleton.md fe73b6431ef362d7e7610fe2b14c3a5e4baf8d37 v1.4.0 v2.11.0 shipped
200
+ skills/vps-deploy .claude/skills/vps-deploy/SKILL.md 3c26d832b2817284b796fde4cd8b03ba85e1d6b8 v1.9.0 v1.10.1 shipped
201
+ skills/vps-deploy .claude/skills/vps-deploy/SKILL.md 5befac0f3775e4a10a83ac54b3d0bb7382c90903 v1.0.5 v1.0.9 shipped
202
+ skills/vps-deploy .claude/skills/vps-deploy/SKILL.md 753f21095c820e391e6b7ca82d02a0db5c3cffc7 v2.0.0 v2.6.0 shipped
203
+ skills/vps-deploy .claude/skills/vps-deploy/SKILL.md 874b758e342422a08ec4d1ad7c04e63f488743bb v1.0.0 v1.0.4 shipped
204
+ skills/vps-deploy .claude/skills/vps-deploy/SKILL.md 9c50315ecd0708fbb638aa36fe008b912edb026c v1.4.0 v1.8.0 shipped
205
+ skills/vps-deploy .claude/skills/vps-deploy/SKILL.md a89ec00b3cfa6c20528a4e9bf4e56534e707e705 v1.1.0 v1.3.0 shipped
206
+ skills/vps-deploy .claude/skills/vps-deploy/references/proxy-ssl.md 00792f9abfd01309bbf02b2bb547314544e9a47c v1.9.0 v2.6.0 shipped
207
+ skills/vps-deploy .claude/skills/vps-deploy/references/rollback.md c3d714bf8b60981e2fe7962dc6827d5b98ab12cb v1.4.0 v2.6.0 shipped
208
+ skills/vps-deploy .claude/skills/vps-deploy/references/scripts.md 5c9b03c897979bc304f73992616eb78f420c6576 v1.4.0 v2.6.0 shipped
package/kit/settings.json CHANGED
@@ -43,21 +43,35 @@
43
43
  "hooks": [
44
44
  {
45
45
  "type": "command",
46
+ "shell": "bash",
46
47
  "command": "cd \"$CLAUDE_PROJECT_DIR\" 2>/dev/null; bash .claude/hooks/guard-bash.sh",
47
48
  "timeout": 60
48
49
  },
49
50
  {
50
51
  "type": "command",
52
+ "shell": "bash",
51
53
  "command": "cd \"$CLAUDE_PROJECT_DIR\" 2>/dev/null; bash .claude/hooks/guard-commit-scan.sh",
52
54
  "timeout": 60
53
55
  }
54
56
  ]
55
57
  },
58
+ {
59
+ "matcher": "PowerShell",
60
+ "hooks": [
61
+ {
62
+ "type": "command",
63
+ "shell": "bash",
64
+ "command": "cd \"$CLAUDE_PROJECT_DIR\" 2>/dev/null; bash .claude/hooks/guard-powershell.sh",
65
+ "timeout": 60
66
+ }
67
+ ]
68
+ },
56
69
  {
57
70
  "matcher": "Write|Edit|NotebookEdit",
58
71
  "hooks": [
59
72
  {
60
73
  "type": "command",
74
+ "shell": "bash",
61
75
  "command": "cd \"$CLAUDE_PROJECT_DIR\" 2>/dev/null; bash .claude/hooks/guard-write.sh",
62
76
  "timeout": 60
63
77
  }
@@ -69,11 +83,13 @@
69
83
  "hooks": [
70
84
  {
71
85
  "type": "command",
86
+ "shell": "bash",
72
87
  "command": "cd \"$CLAUDE_PROJECT_DIR\" 2>/dev/null; bash .claude/hooks/context-usage.sh",
73
88
  "timeout": 60
74
89
  },
75
90
  {
76
91
  "type": "command",
92
+ "shell": "bash",
77
93
  "command": "cd \"$CLAUDE_PROJECT_DIR\" 2>/dev/null; bash .claude/hooks/route-hint.sh",
78
94
  "timeout": 60
79
95
  }
@@ -85,6 +101,7 @@
85
101
  "hooks": [
86
102
  {
87
103
  "type": "command",
104
+ "shell": "bash",
88
105
  "command": "cd \"$CLAUDE_PROJECT_DIR\" 2>/dev/null; bash .claude/hooks/session-guard.sh",
89
106
  "timeout": 60
90
107
  }
@@ -97,6 +114,7 @@
97
114
  "hooks": [
98
115
  {
99
116
  "type": "command",
117
+ "shell": "bash",
100
118
  "command": "cd \"$CLAUDE_PROJECT_DIR\" 2>/dev/null; bash .claude/hooks/session-rehydrate.sh",
101
119
  "timeout": 60
102
120
  }
@@ -107,11 +125,13 @@
107
125
  "hooks": [
108
126
  {
109
127
  "type": "command",
128
+ "shell": "bash",
110
129
  "command": "cd \"$CLAUDE_PROJECT_DIR\" 2>/dev/null; bash .claude/hooks/skill-trust.sh",
111
130
  "timeout": 60
112
131
  },
113
132
  {
114
133
  "type": "command",
134
+ "shell": "bash",
115
135
  "command": "cd \"$CLAUDE_PROJECT_DIR\" 2>/dev/null; bash .claude/hooks/board-sync.sh",
116
136
  "timeout": 60
117
137
  }
@@ -122,6 +142,7 @@
122
142
  "hooks": [
123
143
  {
124
144
  "type": "command",
145
+ "shell": "bash",
125
146
  "command": "cd \"$CLAUDE_PROJECT_DIR\" 2>/dev/null; bash .claude/hooks/session-update-check.sh",
126
147
  "timeout": 60
127
148
  }
package/kit/skills/automode-policy/SKILL.md CHANGED
@@ -60,7 +60,7 @@ Crewforth's). §4.4 needs nothing here: `guard-bash.sh` fails `git commit`/`git
60
60
  not a `permissions.ask` rule — is what asks in the interactive modes. An `ask` rule for those verbs would
61
61
  break the `CLAUDE_GIT_OK` pre-authorisation: a matching ask rule prompts even when a hook returns `allow`.
62
62
 
63
- ## How
63
+ ## How (Bash tool, not PowerShell)
64
64
  ```bash
65
65
  bash .claude/skills/automode-policy/scripts/check.sh # inspect the effective config
66
66
  bash .claude/skills/automode-policy/scripts/apply.sh # propose → diff → ask → install → verify
package/kit/skills/automode-policy/scripts/check.sh CHANGED
@@ -88,13 +88,13 @@ fi
88
88
 
89
89
  if [ "$KITR" -eq 0 ]; then
90
90
  echo " · Crewforth auto-mode rules not in the classifier config (built-ins only) — not a gate either way"
91
- echo " ↳ optional: bash .claude/skills/automode-policy/scripts/apply.sh"
91
+ echo " ↳ optional: bash .claude/skills/automode-policy/scripts/apply.sh (Bash tool, not PowerShell)"
92
92
  exit 3
93
93
  fi
94
94
 
95
95
  if [ "$KITR" -lt 3 ]; then
96
96
  echo " · Crewforth auto-mode rules PARTIAL ($KITR/3 present) — built-ins intact"
97
- echo " ↳ re-apply: bash .claude/skills/automode-policy/scripts/apply.sh"
97
+ echo " ↳ re-apply: bash .claude/skills/automode-policy/scripts/apply.sh (Bash tool, not PowerShell)"
98
98
  exit 3
99
99
  fi
100
100
 
package/kit/skills/crew-board/SKILL.md CHANGED
@@ -10,7 +10,7 @@ metadata:
10
10
  # /crew-board
11
11
  Argument: $ARGUMENTS
12
12
 
13
- Apply the `teamboard` skill. The engine is `bash .claude/hooks/board.sh`; never hand-edit the board.
13
+ Apply the `teamboard` skill. The engine is `bash .claude/hooks/board.sh` (Bash tool, not PowerShell); never hand-edit the board.
14
14
 
15
15
  **Never make the user learn the syntax.** They should be able to run `/crew-board` and answer questions from
16
16
  there; `claim 3` and `done 3` are yours to type, not theirs. Every step below ends in a question with explicit
package/kit/skills/crew-doctor/SKILL.md CHANGED
@@ -6,7 +6,7 @@ metadata:
6
6
  ---
7
7
  # /crew-doctor
8
8
  Verify Crewforth is actually *active* in this project (not just present on disk):
9
- 1. Run `bash .claude/eval/doctor.sh`.
9
+ 1. Run `bash .claude/eval/doctor.sh` (Bash tool, not PowerShell).
10
10
  2. Read its report. It checks: VERSION present · every hook executable · the required git hooks (pre-commit,
11
11
  commit-msg) present · **guard-bash actually blocks a force-push** (catches a hook that is present but neutered) ·
12
12
  `core.hooksPath` points at `.claude/hooks` (else the §4.1/§4.2 commit trace + secret/bloat scan never runs) ·
@@ -33,14 +33,22 @@ with one exception: the skill listing, which a plugin cannot size for itself (it
33
33
  `LC_ALL=C awk -f "${CLAUDE_PLUGIN_ROOT}/eval/lib/skill-listing.awk" "${CLAUDE_PLUGIN_ROOT}"/skills/*/SKILL.md`
34
34
  (prints characters, then skills). At the default fraction a 200,000-token model has 8,000 characters for the whole
35
35
  listing, and Claude Code's own skills take about 5,900 of them (measured on v2.1.282). If Crewforth's number plus
36
- 5,900 is over 8,000, tell the user plainly: on such a model the least-used skills lose their descriptions and stop
37
- being picked; one line in `~/.claude/settings.json` fixes it — `"skillListingBudgetFraction": 0.04`; the cost is
36
+ 5,900 is over 8,000, tell the user plainly: on such a model the least-used skills lose their descriptions and are
37
+ less likely to be picked on their own; one line in `~/.claude/settings.json` fixes it — `"skillListingBudgetFraction": 0.04`; the cost is
38
38
  that the listing is then sent whole every turn, about (that total ÷ 4) tokens, (tokens ÷ 2,000)% of a 200k window.
39
39
  It is their call — show the line and the cost, do not edit their settings. On a 1,000,000-token model the default
40
40
  already fits.
41
41
 
42
+ **On Windows, check Git Bash too.** Without a Git Bash where Claude Code looks, it runs every hook through
43
+ PowerShell and Crewforth's gates do not run. The full install's doctor runs the same script:
44
+ `bash "${CLAUDE_PLUGIN_ROOT}/eval/lib/git-bash.sh"` (Bash tool, not PowerShell). Pass its lines to the user verbatim;
45
+ a ❌ is a failure with its fix. If this session has **no Bash tool at all**, that is the finding itself — Claude Code
46
+ found no Git Bash: tell the user the gates do not run and that the fix is Git for Windows in its default folder, or
47
+ `CLAUDE_CODE_GIT_BASH_PATH` set to its `bin\bash.exe`. Do not run the script through PowerShell.
48
+
42
49
  **The eval scripts are installer-only, by decision.** `eval/` — `doctor.sh`, `smoke-test.sh`, `routing-eval.sh`,
43
- `scan-skill.sh`, `utilization.sh` — ships with `start.sh` and `adopt.sh` and NOT with the plugin edition. The eval
50
+ `scan-skill.sh`, `utilization.sh` — ships with `start.sh` and `adopt.sh` and NOT with the plugin edition; `eval/lib/`
51
+ (the JSON reader, the listing counter, the Git Bash check) ships with both. The eval
44
52
  scripts are developer instruments: they inspect an installation from outside it, and the plugin edition has no
45
- installation to inspect. Say this plainly when someone asks why `/crew-doctor` reports only the listing on a plugin install,
53
+ installation to inspect. Say this plainly when someone asks why `/crew-doctor` reports only the listing (and, on Windows, Git Bash) on a plugin install,
46
54
  rather than treating it as a defect. The Studio panel ships in both editions (`/crew-studio`).
package/kit/skills/crew-gates/SKILL.md CHANGED
@@ -7,7 +7,8 @@ metadata:
7
7
  kind: command
8
8
  ---
9
9
 
10
- Run the report and read it back to the user:
10
+ Run the report with the Bash tool, not PowerShell, and read it back to the user. The report is a file-install tool:
11
+ a plugin install carries no `eval/` commands (only `eval/lib`), by design — there, say so plainly and stop.
11
12
 
12
13
  ```bash
13
14
  bash .claude/eval/gate-report.sh $ARGUMENTS
package/kit/skills/crew-handoff/SKILL.md CHANGED
@@ -6,7 +6,7 @@ metadata:
6
6
  ---
7
7
  # /crew-handoff
8
8
  applies the `handoff` skill:
9
- 1. Read the measured fill: the `🔋 Session` line, or `bash .claude/hooks/context-usage.sh --verbose`.
9
+ 1. Read the measured fill: the `🔋 Session` line, or `bash .claude/hooks/context-usage.sh --verbose` (Bash tool, not PowerShell).
10
10
  2. Write an actionable handover to `docs/SESSION_STATE.md`: Done · In progress (exactly where) · Next step · Open decisions · File pointers · Blockers.
11
11
  3. Preserve decision rationale; the next session should not start from scratch.
12
12
  4. **Cold-reader pass** (`handoff/references/cold-reader.md`): the questions are written from the work BEFORE the
package/kit/skills/crew-skill/SKILL.md CHANGED
@@ -47,7 +47,7 @@ Skipping one of these is how a component ships half-installed:
47
47
  counts it; `/crew-doctor` reports the same number). Raise it in the same commit with the justification comment the
48
48
  file's convention requires. Never raise it to make a red gate green without saying why.
49
49
 
50
- ## 4. Prove it — run all four, in this order
50
+ ## 4. Prove it — run all four, in this order, with the Bash tool, not PowerShell
51
51
  ```
52
52
  bash .claude/eval/scan-skill.sh .claude/skills/<name> # supply-chain: SAFE, and rc=3 means NOT scanned
53
53
  bash .claude/eval/routing-eval.sh # the golden case, positive and negative
@@ -55,7 +55,7 @@ bash .claude/eval/smoke-test.sh # structure, budget, §
55
55
  CREW_NO_STAR=1 bash .claude/eval/doctor.sh # the live install still healthy (star line off here)
56
56
  ```
57
57
 
58
- In the plugin edition none of the four commands above exist — it ships no `eval/` at all — so there the proof is
58
+ In the plugin edition none of the four commands above exist — it ships no `eval/` scripts — so there the proof is
59
59
  the Crewforth repository's own suite, not a local run.
60
60
 
61
61
  **Do not finish while §3b is red.** "It works when I invoke it by name" is not the claim being tested — the claim
package/kit/skills/crew-studio/SKILL.md CHANGED
@@ -47,11 +47,12 @@ fails, say which one and stop — do not improvise a different launch.
47
47
  works the same in both editions.
48
48
 
49
49
  2. **Resolve a runtime — do not ask whether a name resolves.** `ensure-node.sh` sits one level ABOVE the panel
50
- — beside the `server/` directory, not inside it. Both spellings, in full, so there is nothing to guess:
50
+ — beside the `server/` directory, not inside it. Both spellings, in full, so there is nothing to guess (run it
51
+ with the Bash tool, not PowerShell):
51
52
 
52
53
  ```bash
53
54
  bash .claude/studio/ensure-node.sh --explain # full install
54
- bash ${CLAUDE_PLUGIN_ROOT}/studio/ensure-node.sh --explain # plugin edition (step 1 decided which)
55
+ bash "${CLAUDE_PLUGIN_ROOT}/studio/ensure-node.sh" --explain # plugin edition (step 1 decided which)
55
56
  ```
56
57
 
57
58
  It prints the absolute path of a Node 18+ that actually runs, and exits 1 when there is none. Use that path;
package/kit/skills/crew-update/SKILL.md CHANGED
@@ -13,7 +13,7 @@ Bring the Crewforth install up to the latest published version:
13
13
  3. **Latest version:** `npm view crewforth version` (needs network). If it can't be reached,
14
14
  say so and stop — don't guess.
15
15
  4. **Compare.** Already on the latest → report "up to date (vX)" and stop. Otherwise show **old → new**.
16
- 5. **Before:** run `bash .claude/eval/update-guard.sh pre`. It records what is on disk (in `.claude/.state/`, nothing
16
+ 5. **Before:** run `bash .claude/eval/update-guard.sh pre` (Bash tool, not PowerShell). It records what is on disk (in `.claude/.state/`, nothing
17
17
  else) and prints one of three verdicts:
18
18
  - `clean` → nothing to report, go on.
19
19
  - `UNCOMMITTED:` with the changed paths in `.claude/` and `CLAUDE.md` → show the user that list and **ask** —
@@ -30,9 +30,9 @@ Bring the Crewforth install up to the latest published version:
30
30
  to completion instead of blocking on a prompt your shell can't answer.
31
31
  If you'd rather review each handover decision yourself, tell the user to run `npx crewforth@latest update`
32
32
  (no flags) in **their own terminal**, where both the npx and the interactive prompts work.
33
- 7. **After:** run `bash .claude/eval/update-guard.sh post` and show the user its lists — added, changed, moved,
33
+ 7. **After:** run `bash .claude/eval/update-guard.sh post` (Bash tool, not PowerShell) and show the user its lists — added, changed, moved,
34
34
  removed.
35
- 8. **Verify:** run `/crew-doctor` (or `bash .claude/eval/doctor.sh`) so a bad/partial update surfaces immediately.
35
+ 8. **Verify:** run `/crew-doctor` (or `bash .claude/eval/doctor.sh` — Bash tool, not PowerShell) so a bad/partial update surfaces immediately.
36
36
  If the updater's or doctor's output contains a line starting with ⭐, pass it to the user verbatim as the LAST
37
37
  line of your summary — it prints once per Crewforth version, and it is meant for the user, not for you.
38
38
  9. **Report** old → new + the headline changes. Take them from `.claude/.state/whats-new.md` — the updater writes
package/kit/skills/frontend/SKILL.md CHANGED
@@ -8,11 +8,35 @@ description: |
8
8
  # Frontend Discipline (stack-agnostic)
9
9
 
10
10
  <!-- routing-eval reads the next line; why it sits in the body: AGENT_TEMPLATE.md -->
11
- Trigger phrases: "frontend", "screen size", "component", "page", "UI", "state management", "interface"
11
+ Trigger phrases: "in the frontend", "frontend code", "frontend'de", "screen size", "component", "page", "the UI", "UI component", "UI bug", "state management", "interface"
12
12
 
13
- Web (React/Next/Vue/Svelte/Angular), mobile (React Native/Flutter) or desktop — shared principles.
13
+ Web, mobile (native iOS/Android, Flutter, React Native, .NET MAUI, KMP) or desktop — shared principles.
14
14
  The stack-specific "how" (native bridge, router choice, etc.) lives in the relevant project skill; this skill applies to all of them.
15
15
 
16
+ ## Client stack — resolve it before the first line, in this order; stop at the first answer
17
+ There is no default client stack, and none is suggested: the user chooses. **Model discipline, not a gate.**
18
+ 1. **The request says it.** "Add a Flutter screen" is an answer. Use it; record it if `## Stack` was empty.
19
+ 2. **`CLAUDE.md ## Stack` → `Client:` says it.** A filled line is a decision: follow it.
20
+ 3. **The repo says it** — manifests at the root and one level down (`app/`, `mobile/`, `web/`, `client/`, `apps/*`):
21
+
22
+ | Manifest | Client |
23
+ |:--|:--|
24
+ | `package.json` | web (framework from its dependencies); React Native/Expo if it depends on `react-native` or `expo` |
25
+ | `pubspec.yaml` | Flutter |
26
+ | `*.xcodeproj` / `Package.swift` | native iOS (Swift / SwiftUI) |
27
+ | `build.gradle(.kts)` with the Android plugin | native Android (Kotlin / Compose); KMP if it applies `kotlin("multiplatform")` |
28
+ | `*.csproj` with `<UseMaui>` | .NET MAUI |
29
+ 4. **Still open (a new project) → ask, once.** A few questions at most (which platforms · which client stack), the
30
+ options listed plainly — **no option marked recommended and no "Decide for me"**: the stack is the user's call.
31
+ This is the one decision where the discipline's "give a clear recommendation" does NOT apply: state each option's
32
+ trade-off and stop there. That Crewforth ships a layer for one stack (`frontend-rn-expo`) is not a reason to
33
+ favour it (measured: the first version of this step still got "React Native + Expo (recommended)", because of it).
34
+ No one to ask (headless) → write the questions as text and stop; do not start client code on a guessed stack.
35
+
36
+ **Record it** in the `Client:` line of `CLAUDE.md ## Stack`, plus an ADR (`adr` skill). **Never ask twice** — a
37
+ recorded client stack changes only when the user asks. Then apply this skill on that stack; an RN/Expo project
38
+ also applies `frontend-rn-expo`.
39
+
16
40
  ## Architecture
17
41
  - **Presentation / logic separation:** component/view is pure and thin; business logic lives in the hook/composable/service layer.
18
42
  - **Reusability:** repeated UI is factored out; the prop contract is clear and typed.
package/kit/skills/frontend-rn-expo/SKILL.md CHANGED
@@ -7,6 +7,9 @@ description: |
7
7
 
8
8
  # React Native + Expo (stack-specific layer)
9
9
 
10
+ **Apply only when the project IS React Native / Expo** — its `package.json` depends on `react-native` or `expo`.
11
+ Otherwise do not use this skill: `frontend` and the project's own stack apply (see its Client stack order).
12
+
10
13
  <!-- routing-eval reads the next line; why it sits in the body: AGENT_TEMPLATE.md -->
11
14
  Trigger phrases: "expo", "react native", "native bridge", "expo router", "rn screen", "prebuild"
12
15
 
package/kit/skills/handoff/SKILL.md CHANGED
@@ -31,10 +31,10 @@ Session fill (the `🔋` line) > 75% · phase closure · topic change. Goal: the
31
31
  ```
32
32
 
33
33
  ## Measure the session before summarising it
34
- Run **`bash .claude/hooks/session-stats.sh`** first. What it reports belongs in the handover as fact, not
34
+ Run **`bash .claude/hooks/session-stats.sh`** first, with the Bash tool, not PowerShell. What it reports belongs in the handover as fact, not
35
35
  impression — an auto-compaction means state was already dropped before you started writing (say what was lost),
36
36
  a runaway loop marks an approach the next session should not walk back into, and repeated prompts mark context
37
- that never landed and has to be written down explicitly this time. Missing script (plugin install) → say so.
37
+ that never landed and has to be written down explicitly this time. Missing script → say so.
38
38
 
39
39
  ## Principles
40
40
  - **Action-oriented:** focused not on "what was done" but on "exactly where to resume now."
package/kit/skills/reflect/SKILL.md CHANGED
@@ -17,7 +17,7 @@ drives a change *to* its exit test; reflect asks whether the exit test — and t
17
17
  the right one. Skip it for a one-line, unambiguous change; there's nothing to reflect on.
18
18
 
19
19
  ## Measure first, then remember
20
- Run **`bash .claude/hooks/session-stats.sh`** before answering anything below, and open the pass with what it
20
+ Run **`bash .claude/hooks/session-stats.sh`** (Bash tool, not PowerShell) before answering anything below, and open the pass with what it
21
21
  reports. A retro built only on recall is an interview with the least reliable witness in the room: the model
22
22
  reconstructs a tidy story from a context that has already been summarised, and the stretches where it span on a
23
23
  failing approach are exactly the ones it remembers least. The script counts what actually happened — prompts,
@@ -26,7 +26,7 @@ tool calls, failing loops, near-duplicate prompts, interrupts, auto-compactions.
26
26
  Treat each ⚠️ as a **question to answer in the pass**, not a verdict: a runaway loop asks *what assumption kept
27
27
  failing*; a repeated prompt asks *what context never landed*; an interrupt asks *where intent diverged*; an auto
28
28
  compaction asks *what state was silently dropped*. If the numbers and your recollection disagree, the numbers are
29
- the record. If the script is missing (a plugin install has no `.claude/hooks/`), say the retro is recall-based —
29
+ the record. If the script is missing, say the retro is recall-based —
30
30
  do not present recall as measurement.
31
31
 
32
32
  ## The pass
package/kit/skills/teamboard/SKILL.md CHANGED
@@ -18,7 +18,7 @@ that Ali started item #1 two hours ago exists nowhere the other sessions can see
18
18
  a blocked item gets picked up before its dependency lands; whoever inherits half-finished work rebuilds the
19
19
  context from scratch. The board makes those three facts shared.
20
20
 
21
- ## The engine (all commands go through it)
21
+ ## The engine (all commands go through it, with the Bash tool, not PowerShell)
22
22
  `bash .claude/hooks/board.sh <cmd>` — `status` · `show <id>` · `claim <id>` · `done <id> [note]` ·
23
23
  `drop <id> <note>` · `note <id> <text>` · `add <id> <title> [deps] [external]` · `init` · `sync`.
24
24
  Never hand-edit the board; every write must go through the engine or it loses its atomicity.
package/kit/skills/token-budget/SKILL.md CHANGED
@@ -37,14 +37,14 @@ for **isolation / parallelism / a clean window**, or when the isolated work woul
37
37
  6. **Least tooling.** An agent holds only the tools it needs; extras accidentally pollute the context + burn the limit.
38
38
  7. **Lean SKILL.md.** Skills load into the main context; heavy reference goes to a separate file, only when needed.
39
39
  8. **Targeted reading.** Instead of reading a whole file, pinpoint with Grep/Glob.
40
- 9. **Measure fill, don't guess.** `bash .claude/hooks/context-usage.sh` reads it from the transcript; crew-session-manager recommends continue or handoff+clear; the user runs `/clear` at a phase boundary.
40
+ 9. **Measure fill, don't guess.** `bash .claude/hooks/context-usage.sh` (Bash tool, not PowerShell) reads it from the transcript; crew-session-manager recommends continue or handoff+clear; the user runs `/clear` at a phase boundary.
41
41
  10. **Bound what a command hands back.** All of the rules above govern the context's own footprint; none of them
42
42
  govern what a single `Bash` call dumps into it. A `find` over a monorepo, an unfiltered log, a full test run —
43
43
  each returns everything to the main thread whether or not any of it is read. Ask for the answer, not the
44
44
  corpus: `grep -c` over `grep`, `| tail -20` over the whole file, `--quiet`/`--porcelain` where the tool has
45
45
  one, and a redirect to a file plus a pointer when the output is genuinely needed later (rule 2).
46
46
  11. **Cut what nothing reaches — with evidence, not a hunch.** Every installed skill spends its name and
47
- description in EVERY session, forever. `bash .claude/eval/utilization.sh` reports which skills actually fired
47
+ description in EVERY session, forever. `bash .claude/eval/utilization.sh` (Bash tool, not PowerShell; a file install only) reports which skills actually fired
48
48
  in this project's transcripts and how many bytes the cold ones cost, which is the list `doctor.sh` §4a's
49
49
  `skillOverrides: name-only` advice needs and never had. Read it as evidence, not a verdict: a skill that only
50
50
  fires during an incident is doing its job by existing. `--all-projects` widens the scope; by default it reads
package/kit/skills/trace-scan/SKILL.md CHANGED
@@ -18,7 +18,8 @@ a trace leaks sooner or later; this skill + the hooks stop the leak at commit ti
18
18
  - Before crew-commit-agent proposes a message (manual verification).
19
19
 
20
20
  ## How
21
- Pattern list: `./.claude/hooks/trace-blocklist.txt` (grep -iE, one pattern per line).
21
+ Pattern list: `./.claude/hooks/trace-blocklist.txt` (grep -iE, one pattern per line); in the plugin edition
22
+ `${CLAUDE_PLUGIN_ROOT}/hooks/trace-blocklist.txt`.
22
23
  - **Defaults are high-hit:** co-author trailers, auto-generation footers, robot emoji, and AI-assistant/tool
23
24
  brand names. Standalone words that occur too often (model/assistant) are DELIBERATELY excluded. See trace-blocklist.txt for the exact list.
24
25
  - **Vendor name is project-specific:** ADD the name of the third-party template in use to the list (§4.2).
@@ -26,10 +27,10 @@ Pattern list: `./.claude/hooks/trace-blocklist.txt` (grep -iE, one pattern per l
26
27
  Manual scan (a quick look without the hook):
27
28
  ```bash
28
29
  git diff --cached --unified=0 | grep -E '^\+' | grep -Ev '^\+\+\+' \
29
- | grep -iEf .claude/hooks/trace-blocklist.txt
30
+ | grep -iEf .claude/hooks/trace-blocklist.txt # full install; plugin edition: "${CLAUDE_PLUGIN_ROOT}/hooks/trace-blocklist.txt"
30
31
  ```
31
32
 
32
- ## Hook setup
33
+ ## Hook setup (full install only — a plugin cannot set `core.hooksPath`)
33
34
  `start.sh` sets `git config core.hooksPath .claude/hooks` (if there is a git repo). The hooks live
34
35
  under `.claude` → they are in gitignore and stay local (§4.3). To do it later for a repo:
35
36
  ```bash