…ed import

The API server platform never passed fallback_model to AIAgent(),
so the fallback provider chain was always empty for requests through
the OpenAI-compatible endpoint. Load it via GatewayApp._load_fallback_model()
to match the behavior of Telegram/Discord/Slack platforms.

The _session_expiry_watcher retried failed memory flushes forever
because exceptions were caught at debug level without setting
memory_flushed=True. Expired sessions with transient failures
(rate limits, network errors) would retry every 5 minutes
indefinitely, burning API quota and blocking gateway message
processing via 429 rate limit cascades.

Observed case: a March 19 session retried 28+ times over ~17 days,
causing repeated 429 errors that made Telegram unresponsive.

Add a per-session failure counter (_flush_failures) that gives up
after 3 consecutive attempts and marks the session as flushed to
break the loop.

…5046)

When an agent was actively processing a message, /status sent via Telegram
(or any gateway) was queued as a pending interrupt instead of being dispatched
immediately. The base platform adapter's handle_message() only had special-case
bypass logic for /approve and /deny, so /status fell through to the default
interrupt path and was never processed as a system command.

Apply the same bypass pattern used by /approve//deny: detect cmd == 'status'
inside the active-session guard, dispatch directly to the message handler, and
send the response without touching session lifecycle or interrupt state.

Adds a regression test that verifies /status is dispatched and responded to
immediately even when _active_sessions contains an entry for the session.

…kill

launchctl load/unload/start/stop are deprecated on macOS since 10.10
and fail silently on modern versions. This replaces them with the
current equivalents:

- load -> bootstrap gui/<uid> <plist>
- unload -> bootout gui/<uid>/<label>
- start -> kickstart gui/<uid>/<label>
- stop -> kill SIGTERM gui/<uid>/<label>

Adds _launchd_domain() helper returning the gui/<uid> target domain.
Updates test assertions to match the new command signatures.

Fixes #4820

Replace the two-step stop/start restart with a single
launchctl kickstart -k call. When the gateway triggers a
restart from inside its own process tree, the old stop
command kills the shell before the start half is reached.
kickstart -k lets launchd handle the kill+restart atomically.

Telegram polling can inherit a stale webhook registration when a deployment
switches transport modes, which leaves getUpdates idle even though the gateway
starts cleanly. Outbound send also treats Telegram retry_after responses as
terminal errors, so brief flood control can drop tool progress and replies.

Constraint: Keep the PR narrowly scoped to upstream/main Telegram adapter behavior
Rejected: Port OpenClaw's broader polling supervisor and offset persistence | too broad for an isolated fix PR
Confidence: high
Scope-risk: narrow
Reversibility: clean
Directive: Polling mode should clear webhook state before starting getUpdates, and send-path retry logic must distinguish flood control from timeouts
Tested: uv run --extra dev pytest tests/gateway/test_telegram_* -q
Not-tested: Live Telegram webhook-to-polling migration and real Bot API 429 behavior

…n keys

Telegram's Bot API disallows hyphens in command names, so
_build_telegram_menu registers /claude-code as /claude_code. When the
user taps it from autocomplete, the gateway dispatch did a direct
lookup against skill_cmds (keyed on the hyphenated form) and missed,
silently falling through to the LLM as plain text. The model would
then typically call delegate_task, spawning a Hermes subagent instead
of invoking the intended skill.

Normalize underscores to hyphens in skill and plugin command lookup,
matching the existing pattern in _check_unavailable_skill.

…e LLM

Previously, typing a /command that isn't a built-in, plugin, or skill
would silently fall through to the LLM as plain text. The model often
interprets it as a loose instruction and invents unrelated tool calls —
e.g. a stray /claude_code slipped through and the model fabricated a
delegate_task invocation that got stuck in an OAuth loop.

Now we check GATEWAY_KNOWN_COMMANDS after the skill / plugin /
unavailable-skill lookups and return an actionable message pointing the
user at /commands. The user gets feedback, and the agent doesn't waste
a round-trip guessing what /foo-bar was supposed to mean.

The warning said 'forwarding as plain text' but the code returns a
user-facing error reply instead of forwarding. Describe what actually
happens.

- Fix GatewayApp → GatewayRunner import in api_server.py (PR #4976)
- Update launchd test assertions for new bootstrap/bootout/kickstart commands (PR #4892)
- Add nonlocal message declaration in run_sync() to fix UnboundLocalError (pre-existing scoping bug)