a6a57bd5f5
fix(ce-compound): run session-history probe off Claude via SKILL_DIR anchor
...
The automatic Full-mode session-history probe used ${CLAUDE_SKILL_DIR}
guards, which are empty on native Codex/Cursor installs, so the probe
silently took the skip branch there despite being documented as
cross-harness. Migrate the three session-history script invocations
(script-resolution prose, discovery pipeline, extraction pipeline) to
the model-filled SKILL_DIR anchor per AGENTS.md tier-3 guidance.
Address PR review feedback (#1103 )
2026-07-11 01:35:57 -07:00
924cfea3a2
fix(ce-compound): auto-select mode and probe session history
...
Interactive ce-compound asked the user two questions they can't
usefully answer: Full vs Lightweight, and whether to search session
history. Mode depends on remaining context budget the agent can
observe but the user can't; session-history value depends on unrelated
prior sessions neither party knows about a priori. Both are now
resolved by the agent instead of a prompt.
- Full vs Lightweight is auto-selected (default Full; Lightweight only
under real context pressure) and the choice is reported on the first
output line.
- Session history runs as an automatic two-stage probe in Full mode:
the cheap discovery+metadata pass always runs in parallel, escalating
to extraction+synthesis only when a candidate clears a relevance bar
(branch match or >=2 keyword hits).
- Dropped the "What's next?" menu; cross-doc maintenance stays routed to
ce-compound-refresh via the refresh recommendation line.
The only surviving interactive prompt is the discoverability consent,
which edits a tracked instruction file. Headless mode unchanged.
2026-07-11 01:26:52 -07:00
3519336ecd
feat(cline): add native skills install support ( #1033 )
...
Co-authored-by: Trevin Chow <trevin@trevinchow.com >
2026-07-10 09:53:27 -07:00
406ead1d9a
fix(targets): extend user-symlink preservation to codex and opencode writers ( #1094 )
...
Co-authored-by: Claude Code <noreply@anthropic.com >
2026-07-10 09:31:43 -07:00
72c2d16337
fix(ce-test-browser): correct review mode guidance ( #1098 )
2026-07-10 00:45:41 -07:00
ee2fee4c17
feat(ce-plan,ce-brainstorm): dispatch reasoning-heavy steps to Fable on Claude Code ( #1091 )
2026-07-09 17:59:54 -07:00
7aee6fb436
docs(best-practices): capture preserve-user-content completeness learning ( #1090 )
2026-07-09 15:07:56 -07:00
bfa9dd49a2
fix(pi): preserve user-managed symlinks and unmanaged paths on install ( #1089 )
...
Co-authored-by: Claude Fable 5 <noreply@anthropic.com >
2026-07-09 15:04:48 -07:00
f794333aac
fix(ce-commit-push-pr): size PR descriptions by reviewer decision cost ( #1088 )
2026-07-09 14:35:11 -07:00
c5146cf136
refactor(converters): remove dead code in opencode permission builder ( #1060 )
2026-07-09 14:24:32 -07:00
0fffdc7b70
feat: add native Devin CLI plugin support ( #1062 )
...
Co-authored-by: Trevin Chow <trevin@trevinchow.com >
2026-07-09 14:13:52 -07:00
7fd155d317
fix(skills): make pre-resolution commands shell-portable so skills load under PowerShell ( #1078 )
2026-07-09 13:51:02 -07:00
415b472ac8
fix(ce-doc-review): fail fast when document paths are not on disk ( #1079 )
2026-07-09 13:49:33 -07:00
612902260a
fix(models): align plugin behavior with GPT-5.6 ( #1087 )
2026-07-09 13:48:29 -07:00
d9f2569182
fix(ce-plan): recommend ce-work over /goal in post-plan menu ( #1085 )
2026-07-09 10:59:57 -07:00
2cfe3c2b61
fix(grok): support compound-engineering as a native Grok Build plugin ( #1086 )
2026-07-09 10:19:18 -07:00
b498e7a8d1
fix(cli): bump sonnet and opus model aliases to latest generations ( #1083 )
2026-07-08 21:39:16 -07:00
fc0395b8c0
test(ce-explain): guard predict-then-reveal parity across SKILL.md and check-in reference ( #1081 )
2026-07-08 17:05:19 -07:00
494c377280
fix(ce-compound): detect Claude sessions with long metadata preambles in extract-skeleton ( #1082 )
2026-07-08 17:05:01 -07:00
1756c0b9f3
chore: release main ( #1075 )
...
Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
2026-07-08 00:29:49 -07:00
458401e0bb
feat(ce-commit-push-pr): teach new concepts in PR descriptions ( #1077 )
...
Co-authored-by: Claude Fable 5 <noreply@anthropic.com >
2026-07-08 00:28:08 -07:00
b275312619
fix(cli): expand "~/" home shorthand on Windows ( #1055 )
2026-07-08 00:26:23 -07:00
c9e9d62922
fix(tests): bump expected skill count to 29 after ce-sweep landed ( #1065 )
2026-07-07 22:05:23 -07:00
5813075fc3
fix(ce-code-review): call codex via command to bypass shell wrappers ( #1076 )
2026-07-07 22:03:56 -07:00
8559e4cc31
fix(ce-brainstorm,ce-plan): add blindspot pass for unfamiliar territory ( #1068 )
2026-07-07 22:03:37 -07:00
827233cbbf
fix(ce-compound,ce-compound-refresh): validate doc claims at write time ( #1073 )
2026-07-07 21:43:02 -07:00
3a3be15547
fix(ce-code-review): catch silent-pass verification mechanisms ( #1074 )
2026-07-07 21:37:39 -07:00
1287b0985c
chore: release main ( #1059 )
...
Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
2026-07-05 23:21:43 -05:00
2fce3ee246
fix: align skill count with 29 skills on main ( #1067 )
2026-07-05 21:20:42 -07:00
d3f35297ad
feat(ce-sweep): recurring feedback-sweep skill ( #1056 )
2026-07-03 14:45:59 -05:00
be8054c1d1
feat(ce-explain): add personal learning explainer skill with check-in loop ( #1058 )
2026-07-02 23:35:02 -05:00
db21ba21ef
chore: release main ( #1046 )
...
Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
2026-07-01 13:04:20 -07:00
9bc6713c8b
fix(lfg,ce-work,ce-debug): require behavior verification evidence ( #1054 )
2026-07-01 11:43:44 -07:00
90ac7a43fd
fix(ce-plan): require post-review handoff ( #1052 )
2026-07-01 09:50:22 -07:00
6b4e98387c
fix(ce-code-review): review autofix diffs before finishing ( #1049 )
2026-07-01 01:57:19 -07:00
e377cb7a47
fix(ce-commit-push-pr): preserve related work references ( #1051 )
2026-06-30 23:37:54 -07:00
88b47bf2dc
fix(ce-debug): run post-fix polish before PR handoff ( #1050 )
2026-06-30 23:07:38 -07:00
0c5d9ec1f3
fix(ce-code-review): add missed review heuristics ( #1047 )
2026-06-30 18:46:28 -07:00
9ab9809489
fix(ce-commit-push-pr): auto-branch detached PR flows ( #1045 )
2026-06-30 18:41:25 -07:00
ea89ce196f
chore: release main ( #1032 )
...
Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
2026-06-30 17:49:46 -07:00
2bae51b4bb
fix(ce-test-browser): keep pipeline runs from blocking on prompts ( #1044 )
2026-06-30 17:15:08 -07:00
b6729f364d
fix(ce-code-review): defer @-inlined references to on-demand loads ( #1043 )
2026-06-30 17:04:54 -07:00
699f46cca8
refactor(ce-dogfood): single-source the report template ( #1040 )
2026-06-30 16:50:06 -07:00
7aee1800d8
fix(ce-polish): remove stale checklist/Unit-5 sediment from references ( #1041 )
2026-06-30 16:49:12 -07:00
13835643ca
refactor(ce-simplify-code): extract reviewer personas to references/ ( #1042 )
2026-06-30 16:45:49 -07:00
0c5cdfe02d
refactor(ce-plan): single-source the scoping-synthesis gate templates ( #1039 )
2026-06-30 15:59:47 -07:00
888ca42989
refactor(skills): tighten ce-brainstorm and ce-pov from skill evaluations ( #1036 )
2026-06-30 15:55:44 -07:00
8d2f8da8d9
feat(antigravity): enable one-command agy plugin install from GitHub ( #1034 )
2026-06-30 15:49:12 -07:00
f8ff23bfc3
fix(ce-optimize): validate against full schema rule set and bound subagent dispatch ( #1038 )
2026-06-30 15:37:12 -07:00
e5ddd3079b
docs(skill-design): restore inline-vs-extract authoring principles ( #1037 )
2026-06-30 14:47:01 -07:00
Trevin Chow