@lmctl-ai/lmctl 0.1.91 → 0.1.92

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
@@ -2,6 +2,13 @@
2
2
 
3
3
  All notable public-preview changes for `@lmctl-ai/lmctl` are recorded here.
4
4
 
5
+ ## 0.1.92 — 2026-07-08
6
+
7
+ - **Retired `_SKILL_` downloads.** Legacy teamfile `_SKILL_` lines are
8
+ silently ignored for back-compat: `seed` no longer fetches or writes
9
+ `durable-memory/skills/`, and `plan` points at one core Lead skill link
10
+ instead of generating skill-download directives.
11
+
5
12
  ## 0.1.91 — 2026-07-08
6
13
 
7
14
  - **Removed `lmctl wait --id`.** `wait` is an interactive first-return primitive over the caller's queue (id-handshake retired).
@@ -34,8 +41,8 @@ All notable public-preview changes for `@lmctl-ai/lmctl` are recorded here.
34
41
 
35
42
  - **Removed `lmctl wait --all`.** System-wide scope is a foot-gun for a paced
36
43
  wake (field-confirmed): unrelated completions can wake the caller before its
37
- own next job finishes. Scope `wait` with default-self, `--from`, `[teamfile]`,
38
- or `--id`.
44
+ own next job finishes. Scope `wait` with default-self, `--from`, or
45
+ `[teamfile]`.
39
46
 
40
47
  ## 0.1.87 — 2026-07-07
41
48
 
@@ -59,10 +66,9 @@ All notable public-preview changes for `@lmctl-ai/lmctl` are recorded here.
59
66
  - **`wait` is robust against the failure modes of a fire-and-forget loop.** A
60
67
  phantom `running` row whose process died without writing a finish is reclaimed
61
68
  as `error` (dead/stale holder detection) instead of hanging `wait` forever; an
62
- appearance-grace poll covers the `cmd & lmctl wait` insert-lag race; and
63
- `lmctl exec --json` prints its invocation id so `lmctl wait --id <id>` gives a
64
- zero-race handshake. Scopes: default (caller, via `LMCTL_SELF_SESSIONID`),
65
- `--from`, `<teamfile>`, `--all`, `--id`; exit `0` completed/idle, `1` timeout
69
+ appearance-grace poll covers the `cmd & lmctl wait` insert-lag race. Scopes:
70
+ default (caller, via `LMCTL_SELF_SESSIONID`), `--from`, `<teamfile>`; exit
71
+ `0` completed/idle, `1` timeout
66
72
  (8h backstop), `2` usage. A single `idle` return means "nothing running right
67
73
  now", not "done" — loop `wait`.
68
74
 
@@ -113,9 +119,8 @@ All notable public-preview changes for `@lmctl-ai/lmctl` are recorded here.
113
119
  - **MCP install retired.** lmctl no longer writes provider MCP config on
114
120
  seed/chat (the internal `lmctl_chat` code remains; only the installation is
115
121
  removed).
116
- - **Skills over the network.** The default teamfile template pulls shared skills
117
- (background-wakeup, team-lead-workflow, durable-memory) from lmctl.com via
118
- `_SKILL_ <url>` into `durable-memory/skills/` on seed.
122
+ - **Legacy network-skill experiment.** The default teamfile template once pulled
123
+ shared skills from lmctl.com into `durable-memory/skills/` on seed.
119
124
  - **`effort=` advisory for `provider=qwen`.** Qwen Code exposes no reasoning-effort
120
125
  CLI flag, so lint points you to a thinking/effort model id instead of
121
126
  mis-warning that effort is ignored.