- configure proxy: the bilibili_proxy key lost its only reader when yt-dlp
exited the bilibili channel (PR5), leaving docs promising an unlock that
did nothing. Now saves a generic `proxy` key (legacy key kept in sync),
and every wording — CLI output, server tip, install.md quick reference —
says what it really is: a saved address agents read to export
HTTP(S)_PROXY, not an unlock switch
- transcribe: subprocess calls get timeouts (yt-dlp download 1800s, ffmpeg
600s) so a stalled network read can't hang the CLI forever
- docs/update.md: rdt-cli install uses the same pinned git SHA as the
code's _RDT_GIT_SOURCE instead of floating HEAD
- probe.py: documented as side-effect-free-probes-only (retries re-run
verbatim with no backoff)
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
- backends is now an ordered candidate list (first = preferred); channels
report the backend actually serving via active_backend, surfaced in the
doctor text report and --json
- new agent_reach/probe.py really executes upstream commands and tells
apart missing / broken (stale venv shebang after a system Python
upgrade) / timeout, with a reinstall prescription for broken installs
- all 13 channels migrated off which()-only checks: fixes bilibili
false-positive "bili-cli 可用" on broken shims, misleading xiaohongshu
"连接失败", rdt OSError crashing doctor, mcporter breakage masquerading
as "未配置"
- twitter: 15s probe + 1 retry (flaky 10s timeout), broken twitter-cli
now falls back to bird instead of aborting the check
- doctor survives per-channel exceptions; config supports per-channel
backend override (<channel>_backend / <CHANNEL>_BACKEND env)
- fix skill install/uninstall crash on symlinked skill dirs (the
"[Errno None] None" warning from shutil.rmtree on a symlink)
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Pnant