@lmctl-ai/lmctl 0.1.92 → 0.1.94

package/CHANGELOG.md

@@ -2,6 +2,26 @@
 
 All notable public-preview changes for `@lmctl-ai/lmctl` are recorded here.
 
+## 0.1.94 — 2026-07-08
+
+- **Debug diagnostics now write only to the debug log file.** YOLO spawn
+  diagnostics and CLI startup breadcrumbs append to `~/.lmctl/debug-*.log`
+  instead of printing `[debug]` lines to stdout/stderr.
+
+- **Removed `lmctl chat --force`.** The flag defeated terminal-lock and
+  in-flight safeguards by forcing a chat into a session already held by a live
+  process. Dead/stale holders now auto-reclaim without an option; live holders
+  should be reached with liveness-aware `lmctl send`.
+
+## 0.1.93 — 2026-07-08
+
+- **Replaced `--from` with `I_am=<teamfile:alias>`.** Agent-origin commands now
+  use explicit self-identification for `chat`, `send`, `recv`, `wait`, `exec`,
+  and `loop`; the old option is rejected with a hint. `LMCTL_SELF_SESSIONID`
+  is the primary derived identity source, `I_am=` is the explicit fallback when
+  no session marker is available, and a conflicting `I_am=` is rejected so a
+  known session cannot impersonate another member.
+
 ## 0.1.92 — 2026-07-08
 
 - **Retired `_SKILL_` downloads.** Legacy teamfile `_SKILL_` lines are
@@ -41,7 +61,7 @@ All notable public-preview changes for `@lmctl-ai/lmctl` are recorded here.
 
 - **Removed `lmctl wait --all`.** System-wide scope is a foot-gun for a paced
   wake (field-confirmed): unrelated completions can wake the caller before its
-  own next job finishes. Scope `wait` with default-self, `--from`, or
+  own next job finishes. Scope `wait` with default-self, `I_am=`, or
   `[teamfile]`.
 
 ## 0.1.87 — 2026-07-07
@@ -51,7 +71,7 @@ All notable public-preview changes for `@lmctl-ai/lmctl` are recorded here.
   a start row and a finish row; `lmctl wait` blocks on those rows and returns the
   instant the first one in scope finishes (reporting which completed and which
   are still running). `lmctl exec -- <command>` runs any shell command as a
-  tracked invocation, and a backgrounded `lmctl chat … --from "<teamfile>:Lead"`
+  tracked invocation, and a backgrounded `lmctl chat … I_am="<teamfile>:Lead"`
   is tracked the same way — so `cmd & lmctl wait` is the wake for chats and
   arbitrary commands alike. This replaces the fragile "reserve one blocking job
   as the wake" (N−1/1) pattern; all N launches go to the background and `wait` is
@@ -67,7 +87,7 @@ All notable public-preview changes for `@lmctl-ai/lmctl` are recorded here.
   phantom `running` row whose process died without writing a finish is reclaimed
   as `error` (dead/stale holder detection) instead of hanging `wait` forever; an
   appearance-grace poll covers the `cmd & lmctl wait` insert-lag race. Scopes:
-  default (caller, via `LMCTL_SELF_SESSIONID`), `--from`, `<teamfile>`; exit
+  default (caller, via `LMCTL_SELF_SESSIONID`), `I_am=`, `<teamfile>`; exit
   `0` completed/idle, `1` timeout
   (8h backstop), `2` usage. A single `idle` return means "nothing running right
   now", not "done" — loop `wait`.
@@ -112,7 +132,7 @@ All notable public-preview changes for `@lmctl-ai/lmctl` are recorded here.
 
 - **`lmctl terminal` takes a DB lock** (host + pid + heartbeat) so a second chat
   to a member held live in a terminal is rejected instead of racing; a stale/dead
-  or remote lock is reclaimed safely, and `--force` overrides.
+  or remote lock is reclaimed safely.
 - **`--detach` is hidden from the CLI surface.** It stays functional for AI-team
   chat but is the codex/gemini-only fallback; other providers use their native
   tracked-background path.

package/README.md

@@ -65,7 +65,7 @@ Lintable teamfile templates for research, writing, and review teams ship in the
 [documentation](https://lmctl.com/lmctl) to define your first team.
 
 - `lmctl send <teamfile> <alias> "message"` messages a Lead; enqueues if it is up, falls back to chat if down.
-- `lmctl recv [--from <teamfile:alias>]` drains the calling Lead's mailbox consume-on-read.
+- `lmctl recv [I_am=<teamfile:alias>]` drains the calling Lead's mailbox consume-on-read.
 
 ## Model selection