The plan was written when the server had 54 tools in eight families, and
every batch since then changed the facts it states: P10-B added the three
file-write tools, P10-C the input lock, P11 the two session tools and the
protocol work. Read as-is, the plan would send the next person looking for
work that is already done, so the numbers, the family table and the batch
table now describe the 60 tools that actually exist.
The family table was rebuilt from the registration sites rather than from
the previous table: nine families, and the gating column now separates
"appears in tools/list" from "callable", which the old table conflated.
That distinction also fixes the default-config figure - 19 tools, not the
20 stated earlier: get_keylog needs McpKeylog, download_file needs
McpFileTransfer, and list_network_connections lives in the McpMachine
family, so with a default of read-only and every switch off only the
nine server-local read tools, eight read tools on the host and the two
remark/watch tools remain.
Section 5.10's field list for mcp_list_sessions was a guess written
before the tool existed; it now carries the eight fields that D.5
actually returns, with the original guess kept so the change is visible.
The two stale rows in 5.14 and the outdated header comment in 5.13 are
marked done against the commits that did them - bb55f52 bumped the
comment from 53 to 54, 3d388f8 dropped the number altogether when it
added remote_block_input, and every hash cited here was checked to be an
ancestor of HEAD so none of them points at the merged-away history.
Sections 5.1 to 5.10 gained completion markers, with 5.7 marked partial
because verified landed and verify_hint did not.
Everything in this diff is bookkeeping; no code, no behaviour, and no
claim in section 5.11 or the tools/list volume discussion changed.
Co-Authored-By: deepseek-flash
P11 closes the plan's last batch, entirely on the server side: no client
change is required for any of it.
Long-running tools now expose a cancellation flag that every wait loop
checks in 500 ms slices. A cancelled tool returns through the cleanup
branch it already had (late first packets still get absorbed, sub
connections still get closed, pending entries still get deleted), so
there is no second cleanup path to keep in sync, and the host is released
within a second. Three sources set the flag: notifications/cancelled
(matched on the exact JSON-RPC id, with a fallback to the display form so
a request_id copied out of the session list round-trips), mcp_cancel_session
(by request_id, host id or session id), and a failed SSE write when the
client has gone away. The dispatcher rewrites a cancel-induced failure to
-32800 so callers do not mistake it for a host fault.
tools/call requests carrying _meta.progressToken are answered as
text/event-stream on the same POST: the tool runs on its own thread and
both the notifications/progress events and the final JSON-RPC response go
out through one chunked provider. Requests without a token are unchanged.
download_file reports files done vs. expected, upload_file reports bytes
sent vs. total.
Two new local-only tools, mcp_list_sessions and mcp_cancel_session, list
and cancel in-flight sessions (calls, pending one-shots, terminal, remote
control, file transfer, chat). They are deliberately not gated by the
feature switches or readonly mode, since that is the only way out of a
stuck request. The -32003 busy message now names the occupier:
"Device busy: ... (tool=download_file, elapsed=12s)".
power_action gains wait/wait_ms to verify the result against the main
connection's presence (went_offline, came_back, elapsed_ms); not seeing
the transition is reported as ok, not as a failure. kill_process,
operate_window and the four host_control_* tools accept dry_run, which
validates the target against one fresh snapshot and returns what would be
affected without sending anything. Validation rejections now use -32012;
-32008 stays as documented historical alias.
One implementation note worth recording: the per-thread call context uses
FLS, not thread_local. Any static thread_local in this server crashes the
process about a second after start, because some threads come from
shellcode/MemoryModule images that never ran PE TLS initialisation.
Verified against a real Debug server and a real Windows client over MCP:
the smoke test now passes 104 checks with 0 failures, including
cancellation releasing the host in under a second, the cancelled request
returning -32800, notifications/cancelled returning 202, 63 progress
events with exact byte counts for a 64 MB upload, and a cleanly
terminated SSE stream. Not exercised in this pass: power_action's
wait=true path (it would take the test host down) and the -32006
readonly/switch-off section (every switch was on during the run).
Co-Authored-By: deepseek-flash
Mcp_Next_Plan.md lays out what the MCP surface can still gain and how the
existing tools can be improved: a full client-command coverage matrix, a
file-system write batch, a remote-control gap, and a list of protocol-level
blanks. Every conclusion is tied to a file and line so the next step can be
implemented without re-deriving the analysis.
The first draft was written from a read of the code but not verified against
it. Auditing it turned up four claims that would have sent the work in the
wrong direction, and they are corrected here rather than left to be discovered
during implementation:
- The draft wanted a mode=offline for get_keylog because "the tool only sees a
1-15s live window". That premise is false. CKeyboardManager1 builds its
CircularBuffer from the on-disk record file, and SendData replays the whole
buffer from position zero on every (re)connect, so the existing tool already
returns the persisted history whenever kbrecord=Yes. What is actually missing
is only the on/off switch, the clear command, and reporting the state the
client already sends in TOKEN_KEYBOARD_START byte 1.
- Unlike P7, the file-write tools do not plug into OnSystemReady: their ready
packet is TOKEN_DRIVE_LIST, which lands in OnDriveList, where an empty path
is currently read as "list drives" and consumed. Worse, the three finish
tokens have no case at all in the top-level dispatch, so on the MCP path
(hDlg == NULL) they fall into default: and the sub-connection is never
closed. Appendix C now spells out the three wiring layers that every
sub-connection write tool has to touch.
- -32008 carries two meanings, not four: both remaining sites are argument
validation, and the multi-monitor rejection was already moved to -32005.
Dropping the proposed -32013 removes a whole error-code split.
- The draft put server gate state into get_host_detail, duplicating the
switches object that get_server_status already returns; monitors and most
per-host facts are already there too, so capabilities shrinks to the
client's own ability bits.
It also records what the audit added: delete_path inherits the client's
unbounded wsprintf when recursing (the search path got a length guard in
8f718d9, DeleteDirectory and FixedUploadList did not), a single read-only file
fails silently while a directory delete clears the attribute first, and
create_folder means "ensure this path exists" rather than "create the last
component".
Co-Authored-By: deepseek-flash
The server's TLS menu already wrote the trust anchor into the Windows
client's registry and expected a restart. The Linux client grew the same
command handling (166/167) but, like Windows, only read the anchor once at
startup, so enabling or disabling TLS still meant restarting it. Both
handlers now take effect immediately.
TlsChannelLinux.cpp no longer initialises through std::call_once. The
process-level state (context, enabled, load error) lives behind a mutex and
TlsReload() rebuilds it from whatever anchor file is on disk, so
TlsEnabled() reflects the new transport as soon as the anchor is written or
removed. In-flight sessions survive the swap because SSL_new increments the
SSL_CTX reference count; the unit test pins that down by keeping an
established TLS connection open across a reload that frees the old context.
A reload that hits a bad anchor keeps the previous state and only logs: one
bad push should not brick a working client, and the next start is still
fail-closed. The pushed PEM is parsed before it is written for the same
reason -- a corrupt anchor on disk would lock the client out of the server
with no way to correct it remotely.
Only a real transport flip (plaintext <-> TLS) asks for a reconnect. An
anchor swap alone, as in a certificate rotation, leaves the current session
alone and applies to the next connection. The receive thread never closes
the socket itself; it sets g_needReconnect and the main heartbeat loop
disconnects so the outer reconnect loop can re-handshake, matching how the
heartbeat timeout path already works.
IOCPClient now latches the transport per connection (m_useTls) at handshake
time instead of consulting TlsEnabled() on every send and receive. Without
the latch, the second or so between the reload and the reconnect would push
plaintext into an established TLS session, or the reverse, producing
protocol errors for no reason.
RES_TLS (szReserved[19]) is refreshed right before each login, including
the resend after a group change, so the TLS column on the server flips with
the transport instead of showing the value from process start.
The confirmation dialogs now say what each platform does, with translations
in en_US, zh_TW and de_DE; the keys carry the literal \n escape the INI
parser understands.
Ending a sub-connection is not an error. OpenSSL 3 classifies a peer that
closes without close_notify as an unexpected-EOF protocol error, so every
terminal or file transfer session logged an SSL_read failure on the real
servers even though Recv promises 0 for a plain close and neither side sends
close_notify. SSL_OP_IGNORE_UNEXPECTED_EOF restores the intended semantics.
Verified with the reload unit test (RELOAD-OK, covering the in-flight
connection across a context swap, garbage payload rejection and a broken
anchor keeping TLS alive), an abrupt-close test (recv now returns 0 instead
of SOCKET_ERROR) and end to end against a protocol-faithful fake server that
signs MasterSettings with libsign: plaintext login, 166, TLS reconnect with
RES_TLS=1, 167, plaintext reconnect with RES_TLS=0. The Linux client, the
Debug Windows client and the Debug MFC server all build.
The cost is inherent to taking effect immediately: a remote session in
progress is dropped when the transport changes, and the MFC server briefly
shows the client as a duplicate login until the old connection is reaped
and the standby entry is promoted.
Co-Authored-By: deepseek-flash
Co-Authored-By: Claude Code <noreply@anthropic.com>
All of these tools already had a working wire protocol on the Linux
client (SystemManager.h, common/PTYHandler.h, ScreenHandler.h); only the
server-side client-type gates rejected LNX hosts. kill_process,
exec_command, terminal_open/exec/close and remote_open/close/mouse/
keyboard/clipboard now accept LNX and reject only MAC (plus APK where it
applied). No client change, no capability bit.
exec_command and terminal_exec needed platform-aware plumbing because a
POSIX shell understands none of the Windows constructs: the sentinel is
now "PS1=" / the command / "echo __MCP_DONE_<nonce>__$?" over the Linux
PTY, with output decoded as UTF-8. The PS1= line exists because
interactive bash prints its prompt after every line it reads, which
otherwise leaks into stdout; its echo sits ahead of the command echo and
is removed together with it. The Windows ConPTY and legacy non-PTY forms
are unchanged.
The default command whitelist is now platform-specific: identically named
POSIX commands can write files or change system state (find -exec,
sort -o, hostname), so kDefaultCmdWhitelistPosix only lists commands that
stay read-only even with arguments. env is deliberately excluded, since
"env <program>" runs arbitrary programs and would defeat the whitelist
entirely. An explicit McpCmdWhitelist still overrides both platforms. The
forbidden metacharacter set is split the same way: on Linux it covers
& | < > ; backtick and $( / ${ / !, where ! triggers history expansion in
interactive bash. ^ is not a POSIX metacharacter, so read-only commands
like "grep -v ^# /etc/hosts" are accepted on Linux while Windows keeps
rejecting them.
Input injection had to be sent one MSG64 per COMMAND_SCREEN_CONTROL
packet. The protocol allows batching and the Windows client handles it,
but linux/ScreenHandler.h only processes the first message of a packet,
so a batch left key_press stuck down (X11 autorepeat), click
half-pressed and type emitting only its first character. Splitting the
packets is equivalent for Windows, which processes one message at a time
anyway, and keeps already-deployed Linux clients working without an
upgrade.
Verified on a Debug server against a WSL Ubuntu 24.04 client and a
Windows client: exec_command returned real exit codes (0/1/2) with clean
output, -32007/-32008 negatives, and the POSIX-only ^ allowance;
persistent terminals kept cwd and environment, returned 1 for false and
the last exit code for ";" chains, and closed idempotently; kill_process
killed a test process, returned an empty remaining list and skipped
pid 1; remote_open reported 3840x1140, remote_mouse mapped (0.5,0.5) to
(1920,570), xev showed balanced press/release pairs for keys and buttons,
and a Chinese clipboard round-trip came back as UTF-8. The Windows
regression of the same tools was unchanged, including the multi-monitor
gate on remote_open.
Co-Authored-By: deepseek-flash
Both tools used to reject LNX/MAC targets outright. The server now gates
on the target client type instead: a POSIX target must pass an absolute
POSIX remote path without a backslash or a ".." segment (-32602
otherwise), while Windows targets keep their exact previous behaviour.
upload_file needed no new client command; the POSIX receiver already
handles COMMAND_SEND_FILE_V2 and derives the destination from the packet
name. It now also reports a real receiver_ack: the upload session
records dstClientID/transferID, and OnC2CRecvResult matches the 176
acknowledgement from Linux targets, so a destination write failure
becomes -32011 and a missing ack (with bytes in flight) becomes -32001.
download_file needed two pieces on the POSIX side, both shared by Linux
and macOS through common/FileTransferV2.h: a FileCompletePacketV2 per
file with a streaming SHA-256 (new self-contained common/sha256_posix.h,
since BCrypt is Windows-only) so the server can count completed files
and verify digests, and a TOKEN_DOWNLOAD_ERROR reply when the requested
paths do not exist or are denied, which the server maps to
-32009/-32010 within milliseconds instead of waiting for the stall
watchdog.
Zero-byte files were lost when uploading to POSIX targets because the
worker sends no data chunk for them and the complete packet carries no
path, so the receiver had nothing to create. The server now injects a
zero-length final chunk for each empty file, only for POSIX targets;
Windows targets keep their existing behaviour.
Also fixes a download completion edge case found while testing: the
client decides entry order, so a directory may arrive after all files
(an empty subdirectory, for example), and the directory branch only
completed when the directory count alone covered totalFiles. The check
is now filesDone + directoryCount >= totalFiles.
Verified end to end against a WSL Linux client and a Windows client:
upload and download of a tree with a 3 MB binary, a Chinese filename, an
empty file and an empty subdirectory all round-trip with matching
SHA-256; the documented error paths (-32602, -32009, -32011, -32001) and
the Windows upload/download/C2C regressions were exercised as well.
Co-Authored-By: deepseek-flash
The C2C file transfer tool required both ends to be Windows clients. A Linux
client can already land the files: the Windows source worker builds every
entry name by plain string concatenation of dst_dir and the relative path, so
a POSIX dst_dir such as /tmp/recv arrives as "/tmp/recv\sub\file", which the
POSIX receiver normalizes to "/" and writes below an absolute path. What was
missing was the server gate and any way to tell whether the writes succeeded.
The server now accepts LNX/MAC destinations instead of rejecting them, and
validates dst_dir by destination type: an absolute Windows path for Windows
targets, an absolute POSIX path without backslashes for POSIX targets. The
paths and dst_dir arguments also reject embedded NULs, and a POSIX dst_dir is
converted to CP936 exactly so unencodable characters fail loudly instead of
silently becoming '?'.
To make receiver_ack meaningful, a Linux client that receives a C2C transfer
now reports what it wrote. It tracks the outcome per transferID, counting a
directory entry, the last chunk of a file and the completion packet of an
empty file as the three ways an item is settled, with done sets so that
source retries and the trailing completion packet never double count. When
every item is settled it sends one C2CSendResultPacket back on its main
connection under the new TOKEN_C2C_RECV_RESULT (176), with the first errno in
the win32err field and the mapped FEV2_* status. A dedicated token is used
instead of reusing 175 because a host can be the destination of one transfer
while being the source of another, and a destination result must never latch
the source side's done flag.
The server waits up to five seconds for that acknowledgement after the source
reports success, and only for destinations advertising the new
CLIENT_CAP_C2C_RECV_ACK capability bit (Linux only, macOS reports nothing).
A failed acknowledgement becomes -32011 with the status, errno and received
file count; a missing acknowledgement stays an error only for transfers that
actually moved bytes, since a transfer of empty files alone gives the
receiver nothing to learn totalFiles from. Windows destinations are unchanged
and skip the wait entirely.
Verified on the debug instance: Windows to Linux (WSL Ubuntu 24.04) with
subdirectories, an empty directory, a Chinese file name and a 3 MB binary
lands with matching SHA-256 and receiver_ack true; a Windows-style dst_dir
returns -32602; an unwritable POSIX destination returns -32011; the existing
Windows to Windows cases still pass with receiver_ack false and the same
latency.
Co-Authored-By: deepseek-flash
Adds the 54th MCP tool, copy_files_c2c, which asks the source client to
push a set of files straight to another online client. Previously the only
way to move files between two clients was to download them to the server
machine and upload them again, so the data crossed the wire twice and
needed free disk on the server. With C2C the server only relays packets
and never parses or stores the payload.
Protocol. Two new ids in common/commands.h: COMMAND_C2C_SEND_FILES (174),
carrying the destination directory followed by a NUL-separated path list,
and TOKEN_C2C_SEND_RESULT (175), a fixed 45-byte result packet sent back
over the same sub-connection the transfer used. Clients advertise support
with the LOGIN_INFOR capability bit CLIENT_CAP_C2C_PATHS (0x0008), exposed
server side as Context::SupportsC2CPaths(). File transfer V2 already knew
how to push to dstClientID, so the destination side needed no change at
all; existing V2 receive paths keep working, and older clients simply do
not advertise the bit and are rejected with a clear error instead of
hanging.
Client. OnC2CSendFiles expands environment variables, collects directories
recursively, opens its own sub-connection (with sub-connection auth), and
runs FileBatchTransferWorkerV2 in a detached thread, then reports the
result on that same connection. Every early exit now reports a structured
failure rather than staying silent. The worker needs the process-level
InitFileUpload call that the receive path performs lazily, and a source
client may never have opened a file manager or received an upload, so the
initialization is factored into EnsureV2TransferInited() and called from
both paths.
Server. copy_files_c2c mirrors the existing upload tool: gates, id parsing,
capability checks, path validation, then a session keyed on the source host
that WaitC2CDone polls until the result packet arrives. The C2C relay
branch used to attach a CDlgFileSend progress dialog to the source's
sub-connection; NotifyProc then routed every later packet on that
connection to the dialog, which parsed the result packet as a V1 file
chunk and dropped it, so the tool always timed out even though the files
had arrived. IsC2CTransferPending() now suppresses dialog creation for
MCP-driven transfers only, leaving manual Ctrl+V transfers untouched.
Errors map to -32001 for a source-side transfer failure, -32010 for access
denied, -32009 for a missing path and -32002 for an offline destination.
Both hosts must be Windows: the destination is restricted too, because the
source embeds dst_dir in each V2 chunk name as an absolute Windows path and
only the Windows receiver applies it literally. A Linux or macOS receiver
never gets a target directory here (the clipboard flow sets one with
COMMAND_C2C_PREPARE, which this path does not send), so it would strip the
drive letter and write under / instead. Supporting that needs a prepare
packet carrying an explicit path plus an upgrade of those clients, so for
now the tool rejects them with -32005 rather than writing somewhere wrong.
Verified end to end on the Debug build: three files including a non-ASCII
name, an empty directory and a 3 MB binary arrived with matching SHA-256
in 3.2 s, a non-existent destination directory was created, and the
negative cases returned the documented codes. mcp_smoke.ps1 passes 41/41
with two new assertions and 27/27 in -SkipHost mode. Design notes, error
table and limitations are in docs/Mcp_FileTransfer_Design.md section 14
and docs/Mcp_Design.md section 15.
Co-Authored-By: deepseek-flash
List and query tools put only a human summary in content and kept the real
payload in structuredContent. Clients that hand just content to the model
(WorkBuddy does) therefore never saw a host id: search_hosts answered "1 host
matched" and every follow-up tool was out of reach, and guessed ids 1-3100
could never work because host ids are 64-bit snowflake values.
BuildResult now appends the serialized structuredContent as a second text
block, as the MCP spec recommends for tools with structured output. The append
sits in the single result choke point, so every tool in McpServer, McpMachine
and McpChat gains it, including future ones; empty action results such as
remote_mouse keep their bare "ok".
Tools whose text already carries the payload expose only the difference:
exec_command and terminal_exec add exit_code, get_keylog adds counters,
get_activity_history exposes records without the duplicated raw dump, and the
hosts file reader drops its copy of the content. The override rides in a $
key that BuildResult consumes and removes, so responses change only by the
extra block.
Verified on a Debug instance: list_all_hosts, search_hosts, get_server_status
and list_registry return parseable JSON in content[1] with ids intact, and
tools/mcp_smoke.ps1 -SkipHost passes 27/27 with four new assertions.
Co-Authored-By: deepseek-flash
MCP was served by a plain httplib::Server, so remote clients could only use
cleartext HTTP. YAMA now terminates TLS itself: a small front-end owns the
MCP port, sniffs the first byte and either runs a Schannel handshake and
forwards the decrypted bytes or passes plaintext through, with httplib
moving to a loopback-only internal port. No OpenSSL, no frp change, existing
local HTTP clients keep working, and an unusable certificate fails closed
instead of falling back to cleartext.
The identity loader grew the two things a public certificate needs. ECDSA
(P-256/P-384/P-521) keys are persisted by importing PKCS#8 with an
NCRYPTBUFFER_PKCS_KEY_NAME parameter, the only route MS Software KSP
supports for Schannel; RSA keeps its old export-and-repersist path.
Fullchains are accepted, and the intermediates are added to the current
user's CA store so Schannel, which builds chains only from system stores,
actually sends them. The shared handshake also now buffers a full TLS
record before the first AcceptSecurityContext, which a partial ClientHello
would otherwise poison (SEC_E_UNKNOWN_CREDENTIALS).
Session resumption exposed a second trap: when a client resumes a session,
ChangeCipherSpec, Finished and the first application record arrive in one
segment and AcceptSecurityContext leaves the application record pending in
the session. Nothing drained it, because both callers only decrypt when new
ciphertext arrives, so the request waited until the peer gave up (the MCP
front-end saw no request at all and the client reported "fetch failed
(other side closed)"; the device port had the same hole behind its
occasional login timeouts). Decrypt now drains pending data on len == 0, and
both the front-end and the device port's handshake thread use it before
entering their receive loops.
Configuration is McpHttps, McpCertPath, McpKeyPath and McpInternalPort, the
first three editable in the MCP settings dialog (the certificate rows stay
greyed out until HTTPS is enabled and saving rejects missing files).
Verified with a real Let's Encrypt ECDSA fullchain (53 tools over TLS, 401
with no or bad token), a local three-level chain whose intermediate was
absent from the store, Node 24's fetch on a resumed session (initialize,
notifications/initialized and tools/list all complete), and the device
port's RSA handshake as a regression check. Details in docs/Mcp_Design.md
section 13.
Co-Authored-By: deepseek-flash
A sub-connection's first packet doubles as the readiness packet: the file,
service, registry, process, log and terminal tools all get answered that way.
Once a tool has given up (timeout or failure) its pending entry is gone, so if
the client only gets around to sending that first packet afterwards,
MessageHandle cannot tell it apart from a human click and opens the matching
modeless dialog - a File Manager window appeared while timeouts were being
exercised on this machine.
The wait paths now record, per device, which first-packet token the abandoned
tool would have sent (MarkLateFirstPacketForTool from WaitPending,
WaitPendingRegistry, WaitPendingWrite, WaitFileTransferDone and
WaitTerminalReady), and every dialog-opening case consumes that marker before
falling back to the MFC path. The marker is one-shot, expires after five
minutes and is never set on the success path, so normal use is unaffected; the
worst case is that a single manual open on a host that just had an MCP timeout
is swallowed.
ClientMsg packets are dropped with a log line while a tool is pending for that
host: new clients still send the legacy popup message alongside the structured
failure token, which would otherwise put a modal box on the server UI during an
MCP call.
Co-Authored-By: deepseek-flash
Spell out the new token layout and who sends it, add -32010 to the error
table, and narrow -32008 to parameter validation now that capability
limits report -32005. Document McpToolTimeoutMs as an ini-only key and
correct the note about where the list_files 500-entry cap is applied.
Update the remote-control design and the test report so P2, P6 and P8
show as fixed with the measured evidence.
Co-Authored-By: deepseek-flash
Re-verified every finding of the 2026-10-25 MCP test report against the source and
then end-to-end on the Debug instance, and fixed the ones that live on the server.
P1/P3/P4 share one root cause: the sentinel was appended to the command line, so any
command that swallows the rest of its line (rem, cmd /c, if, for) swallowed the
sentinel too and the call always timed out. On ConPTY clients the command and the
sentinel are now sent as two lines, the sentinel prints %errorlevel% (whose echoed
text is a literal template, not digits) so folding echo can no longer be mistaken
for the sentinel, and FindSentinel returns the real exit code (9009, 1, 0, ...)
instead of 0/1. Old non-PTY clients keep the single-line form, since their echo
skipping depends on the write length.
ConPTY also returns a rendered screen stream rather than a byte stream, so stripped
VT sequences leave duplicate text behind; output cleaning now cuts through the
echoed command line, strips the echoed sentinel line and erases any leftover
sentinel text by its per-command nonce. StripAnsi no longer leaves the backslash of
an OSC terminated by ST.
A terminal session destroyed by a timeout used to report the same "not found" text
as a typo in the session id; a 600s tombstone now says what happened instead.
download_file gets a stall watchdog: no data for 60s fails with a clear message and
releases the host, where a missing remote path used to hold the whole host for the
full 600s default timeout.
Also corrects the get_activity_history annotation (it does reach the remote host),
completes the -32007/-32008/-32009 error table, documents the sentinel design in
Mcp_Design.md 12.1, and corrects the report's claim that the tools blocked during a
transfer are local DB reads -- they are remote machine-manager fetches, and purely
server-local tools are unaffected.
Co-Authored-By: deepseek-flash
CChatManager::OnReceive opened with the same heartbeat guard the other
managers use, `if (szBuffer[0] == TOKEN_HEARTBEAT) return;`, before it
looked at the length. TOKEN_HEARTBEAT is 101, which is also the letter e,
and this channel carries free text rather than command bytes, so every
message starting with e -- "echo", "error", anything -- was silently
discarded. The sender never learned: the server's chat dialog had already
written the text into its own log, so the two sides disagreed about what
was said.
The guard is dead code here. Heartbeats travel on the main connection
only, as a multi-byte [101][Heartbeat] packet consumed by the main
dispatcher; the server never sends one on a chat sub-connection. Delete
it and keep the length check that follows.
The server keeps prepending a space when the GBK payload starts with
101, because already-deployed clients still carry the guard and losing
the message outright is worse than an invisible leading space. Note the
same guard, harmless today, still sits in the machine, file, proxy and
screen-hide managers.
Verified with the Debug instance and a local Debug client: with the
server's space temporarily disabled, chat_send "echo test" and
"second line" both showed up verbatim in the client dialog (read back
with WM_GETTEXT); with the space restored the first arrives as
" echo test". Details in docs/Mcp_Chat_Design.md section 5.
Co-Authored-By: deepseek-flash
The suite only exercised the happy paths, so a regression in the
negative paths -- unknown method, unknown tool, a missing required
argument, an oversized batch or a policy refusal -- would have gone
unnoticed. Add those checks alongside a section that reads the switches
from get_server_status, picks a feature that is actually turned off and
asserts its tool answers -32006; with every switch on it skips rather
than guessing, and with -AuditLog it also asserts the refusal shows up
as CALL_DENIED.
Document the full error-code table in section 12, including which codes
mean "retry with different arguments" and which mean "this is a policy
decision, retrying is pointless", and note that a batch call never
fails as a whole for per-host errors.
Verified against the Debug instance with McpKeylog temporarily set to
zero: get_keylog refused with -32006 and the audit log gained the
matching CALL_DENIED line, 34 checks passing; with the switch restored
32 pass, and 37 with a search path.
Co-Authored-By: deepseek-flash
Add four optional arguments to search_file that narrow the match list
after the client has returned it: min_size and max_size in bytes,
modified_after in either YYYY-MM-DD or YYYY-MM-DD HH:MM:SS form, and
name_only to keep just path and name per entry. The client still runs
the same name-matched recursive search, so these filters cut the
response size and the reasoning work on our side, not the scan time on
the host -- the summary sentence says so and totalAll keeps the
pre-filter count so a too-narrow condition is obvious. Invalid values
(negative sizes, an inverted range, a malformed timestamp) are rejected
with -32602 before any packet goes out.
The timestamp validator only accepted the ten-character form because
its shape loop never allowed the two colons, which rejected every
nineteen-character value; the loop now whitelists them.
Smoke tests cover the unchanged totalAll under a size filter, the slim
name_only payload and a rejected timestamp, for 33 checks when the
search path is supplied. Design notes are in section 11.4.
Co-Authored-By: deepseek-flash
list_online_hosts and list_all_hosts could only page, so finding "the
hosts in this group" or "the Linux ones" meant walking pages or falling
back to search_hosts, which has no client type filter and treats online
and offline hosts differently.
Both tools now accept group, os and client_type. They are substring
matches, combined with AND and case-insensitive over ASCII, matching the
semantics search_hosts already used, and the shared implementation keeps
the two tools from drifting apart. The filter runs before paging, so
total counts the matches and totalAll keeps the pre-filter size, and the
summary switches to the "matches" wording with a trailing line naming
the unfiltered size - a mistyped condition then shows up immediately
instead of silently returning nothing.
Offline hosts carry no clientType in their stored record, so a
client_type filter never matches them; the tool description says so and
points at os for that case.
Verified on the Debug instance: client_type=exe matches the single EXE
host while client_type=lnx matches nothing and still reports totalAll=1;
list_all_hosts with group=default matches all 3161 recorded hosts and
splits them into 1 online and 3160 offline, and os=zzz matches none. The
smoke test gains two checks and passes 28/28.
Co-Authored-By: deepseek-flash
Answering "did my rebuild actually take effect" or "why was that tool
refused" used to mean reading list_online_hosts and get_audit_log and
inferring the rest. A single read-only call now returns the server's own
state: brand and version, the exe link time (taken from the file's last
write, which unlike __DATE__ reflects the whole build), how long the
process has been up, the address and port MCP is actually listening on,
every MCP switch, online and offline host counts, the number of sessions
in flight per family, and how many source IPs are currently blocked.
The bind and port are recorded by Start rather than re-read from the
configuration, so the tool reports what the listener really got. Session
counters take each subsystem's lock in turn and release it immediately,
so the numbers are an estimate of one moment, not a consistent snapshot.
Nothing sensitive is exposed: the token and the command whitelist never
appear, and the switches are reported as booleans only.
The tool is registered as read-only, non-destructive, idempotent and
server-local, which brings tools/list to 53 entries. The smoke test gains
three checks for it, including one that fails if a token-shaped field
ever appears in the output.
Verified on the Debug instance: the summary line reports version, uptime
and listener correctly, sessions.terminal goes 0 to 1 while a persistent
terminal is open and back to 0 after terminal_close, and the smoke test
passes 26/26.
Co-Authored-By: deepseek-flash
get_audit_log used to hand back every entry in the main window's message
mirror on each call. The mirror is a 1000-entry ring buffer, so a busy
server returned a lot of text the caller never looked at, and there was
no way to ask for just the MCP entries or just the ones mentioning a
particular subject.
The tool now takes the same limit/offset pair as the other list tools
and two optional substring filters: type matches the type column and
keyword matches the message text, both case-insensitive. Filtering runs
before paging, so total counts the matches while totalAll keeps the size
of the whole mirror, and the summary line adds the pre-filter count when
a filter is in use so a typo is easy to spot. The GBK to UTF-8
conversion moved into the filter loop so the comparison happens between
strings in the same encoding, still under the dialog's critical section.
The output schema gains total/totalAll/offset/limit/hasMore, and the
smoke test now asserts the paging fields, the limit bound and an empty
result for a filter that cannot match, bringing it to 23 checks.
Verified on the Debug instance: limit=2 returns entries 1-2 then 3-4
with hasMore set, offset=999 reports the out-of-range text, type=mcp
matches the "MCP命令执行" entry, type=提示 matches both hint types and
keyword=监听 matches the listening-port line.
Co-Authored-By: deepseek-flash
Every policy denial - read-only mode, one of the Mcp* switches turned
off, or a host that lacks the required capability - answers -32006. The
audit log only ever showed CALL_FAIL for those, which lumped them
together with parameter errors, offline hosts and timeouts, so after the
fact there was no way to tell "nobody called this" from "it was blocked
by policy".
The hooks all live in one place: the tools/call exit in the HTTP
handler now also reads error.code and error.message from the response it
already parses, and appends a CALL_DENIED event when the code is -32006.
The original CALL_FAIL line is still written, so its statistics keep
their old meaning, and the reason text is the fixed English sentence
each branch already returns - no tool arguments or user input reach the
log. Tool names taken from the request are flattened to a single line
first, both in the denied event and in the existing CALL_OK/CALL_FAIL
lines, so a crafted name cannot forge extra audit rows.
Also refresh the stale note in the design document that still called the
client search MAX_PATH overflow unfixed; the source was fixed in
8f718d9, but already-deployed clients still carry it, which is why the
smoke test keeps its search cases opt-in.
Verified on the Debug instance: with McpFileTransfer temporarily set to
0, download_file returns -32006 and the audit log gains the expected
CALL_DENIED line right after its CALL_FAIL; with the switch restored the
same call is CALL_OK again and no denied line appears. The read-only
smoke test passes 20/20.
Co-Authored-By: deepseek-flash
A file transfer session holds raw pointers to its file-manager and
stream sub-connections, but nothing on the offline path looked at them.
OfflineProc cleans up terminal, remote-control and chat sessions and
skipped file transfers entirely, so a client that disconnected in the
middle of a transfer left the session pointing at a context that went
back to the connection pool. When the tool thread timed out it called
CancelIO on that stale pointer, which could cancel somebody else's
connection.
Add a context-to-device table next to the sessions, kept under the same
mutex and erased by ClearFileTransfer so a recycled pointer is never
mistaken for a live one. OfflineProc now calls
OnFileTransferContextClosed, which drops both pointers. A dropped
stream sub-connection means the transfer cannot complete, so it also
fails the session with error 1003 and wakes the waiting thread; the
file-manager sub-connection only issues the command and its loss is not
treated as a failure of the transfer itself.
Verified against the debug client: killing it 1.2 s into a 500 MB
download returns "Connection lost: the transfer sub-connection closed"
after about 1 s instead of waiting out timeout_ms, and the server stays
up. After the client restarts, two consecutive downloads and one upload
still work.
Co-Authored-By: deepseek-flash
The completion check compared filesDone against totalFiles minus
directoryCount and only ran on COMMAND_FILE_COMPLETE_V2. The client
sends no COMPLETE for directory entries, so a download of an empty
directory, or of a tree that holds only subdirectories, had zero
expected files and could never satisfy the check. The call sat there
until timeout_ms expired (600 s by default) and then reported 1002.
Mark the session done as soon as the directory entries account for
every entry the client declared. Verified with the debug client: an
empty directory returns in 0.2 s with zero files, a tree of nested
subdirectories returns in 0.2 s with the directories created, and a
directory holding three files still downloads them with the same
sizes and hashes as before.
Co-Authored-By: deepseek-flash
Cap MCP request bodies at 2 MB, set a 15 s idle read timeout and a 60 s
write timeout, and give httplib an explicit 32-thread pool: long tools
such as chat_wait hold a worker for up to 55 s, and the default pool of
8 to 15 left other callers queued.
Every tool returned by tools/list now carries readOnlyHint,
destructiveHint, idempotentHint and openWorldHint. The table is
per-tool; anything not registered defaults to destructive so a new
write tool cannot silently look harmless. Read-only tools are forced
to destructiveHint=false.
get_host_detail appends the technical fields the host store already
holds (cpuCores, memoryGB, systemBits, programBits, user, isAdmin,
installTime, expiredDate, monitors, exeVersion, clientPath, pid,
publicIp). The JSON writer is pinned to 15 significant digits so
memoryGB renders as 27.8 instead of 27.800000000000001. list hosts
output is unchanged.
Add host_set_remark and host_set_watch_level. Both validate the host
against the store, are refused in readonly mode, repaint the affected
row, and are audited without recording the value written.
Add search_file, which drives the client file-manager search channel
and returns paged results. The root path is normalized before sending:
the client enumerates with wsprintf("%s\*.*"), so a trailing
backslash produced C:\*.* and FindFirstFile failed, returning zero
hits instantly. The client only searches recursively, so non-recursive
requests are filtered server-side by comparing the parent directory.
Collection stops at 5000 records or 4 MB and the client is told to
stop; on timeout the partial results are returned. Search result
packets are never closed early: the sub-connection stays open until
TOKEN_SEARCH_FILE_FINISH, because the client's search thread holds a
pointer into the file manager object.
Batch reads report a skipped counter and are bounded to 50 ids and a
50 s deadline.
Add tools/mcp_smoke.ps1, a read-only end-to-end smoke test that asserts
auth rejection, the payload limit, annotation coverage, paging fields
and the main read tools. It takes the token from the environment and
never prints it.
Document all of the above in docs/Mcp_Design.md section 10, including
the client-side MAX_PATH stack overflow that a deep search_file call
can trigger, together with the workaround of searching shallow
directories.
Co-Authored-By: deepseek-flash
Expose hosts the server already knows about. list_all_hosts enumerates
both live contexts and the historical records in the client list, with
authStatus as a three-state string instead of a boolean. search_hosts
now honours online=false instead of returning an empty list, while the
default stays online-only so existing callers see the same results.
get_host_detail falls back to the last known record for an offline host
rather than failing with -32002.
Let read-only tools take an ids array of up to 50 hosts in one call.
The batch path reuses the single-host dispatcher and pages per host, so
behaviour cannot drift; write tools reject ids with -32602 so a mistake
cannot fan out across machines.
Add get_keylog behind its own McpKeylog switch. It drives the client's
existing keylogger sub-connection, accumulates the streamed text for a
bounded window and byte cap, then closes the sub-connection and returns
UTF-8 lines. The client pushes its buffered history first, so a result
contains both history and fresh keystrokes; the tool description says so.
Harden the endpoint. Failed bearer checks are counted per source IP and
ten failures within a minute block the IP for fifteen minutes. Auth
failures and tool calls append to logs\mcp_audit.log with metadata only:
no token, no arguments, no response bodies, and rotation at 5 MB. Loopback
sources are audited but never blocked, because the built-in frp tunnel
makes every remote request arrive from 127.0.0.1 and banning that address
would lock out legitimate clients along with the attacker.
Give list-style tools one pagination contract (limit/offset plus
total/totalAll/hasMore) and fix the generated schemas so id and ids are
genuine alternatives: the stale required:[id] left over from the
single-host schema would otherwise reject ids-only calls.
Co-Authored-By: deepseek-flash
Add chat_open, chat_send, chat_wait and chat_close to the MCP tool set,
behind a new McpChat switch that defaults to off and is independent of
the read-only switch. The tools reuse the existing client chat link: the
main connection sends COMMAND_CHAT, the client opens a sub-connection
with TOKEN_CONN_AUTH and TOKEN_CHAT_START, the server answers with
COMMAND_NEXT_CHAT, and both directions then carry raw CP936 text.
Sessions are per host and held in shared pointers under their own mutex,
so a waiting chat_wait survives a close or a dropped sub-connection, and
chat_send is serialised per session. Text is converted strictly between
UTF-8 and CP936: invalid UTF-8, characters GBK cannot represent (emoji,
accented letters) and messages over the client's 4000-byte buffer are
rejected with -32602 instead of being mangled. A one-second settle delay
after the sub-connection binds keeps the first message from reaching the
client before its chat window exists.
The routing hooks run before the host-management hooks in MessageHandle
and clear the route in OfflineProc before the context leaves the list. A
five-second tombstone absorbs a late TOKEN_CHAT_START after a timeout, so
a failed chat_open does not pop the MFC chat dialog. Idle sessions are
swept on every tool call, and audit entries record the session id and
byte counts only, never message bodies.
The settings dialog gains the chat checkbox and the translated labels
for all three languages, and the strict UTF-8 to CP936 helpers move from
McpMachine.cpp to McpInternal so both tool families can share them.
Co-Authored-By: deepseek-flash
Expose the online-list TLS column through list_online_hosts, search_hosts
and get_host_detail. The value comes from the flag the client reports at
login, so old clients always read false.
Add list_network_connections and kill_network_connection on top of the
connection manager sub-connection protocol. The kill path matches the
original 110-byte record instead of rebuilding addresses, supports a dry
run, refuses UDP/IPv6 targets, and verifies that the quadruple really
disappeared instead of assuming success.
Add a host_* family backed by the machine manager protocol: ten read
tools covering every tab of the host management dialog, and four
recoverable write tools for process freeze/thaw, window restore/hide/
maximize/minimize, service start/stop/pause/continue/restart, and
scheduled task run/stop. Every write re-reads the snapshot before sending
and re-checks afterwards, reporting -32009 for stale targets rather than
acting blindly. The family is gated by McpMachine, the risky subset by
McpMachineDangerous, and writes additionally require read-only mode to
be off.
Route managed sub-connections from MessageHandle and OfflineProc, keep
the single-in-flight rule by sharing the pending table with a request
kind, and move the shared JSON/lock/encoding helpers into McpInternal so
both tool families use one implementation. Settings dialog, design docs
and en_US/zh_TW/de_DE translations are updated accordingly.
Co-Authored-By: deepseek-flash
Add optional one-way TLS to the MFC server (server/2015Remote) so the device
port can serve both TLS clients and legacy plaintext clients on the same port,
mirroring the Go server.
Server: on a connection's first bytes, sniff for a TLS ClientHello (0x16 0x03
+ handshake type 0x01). When [tls] Enabled=1 and the identity loads, hand the
socket to a short-lived blocking handshake thread (Option A) running Schannel
AcceptSecurityContext, then decrypt on recv and encrypt on send; plaintext
connections replay the sniffed bytes straight into the existing parse path, so
the plaintext code path is otherwise unchanged. The identity is loaded once
from [tls] CertPath/KeyPath as PEM + CNG (PKCS#8) — the same key pair the Go
server uses — and a missing [tls] section, an empty value, or a load failure
leaves plaintext serving intact and fail-closes any TLS ClientHello.
Also fix two transport-reliability bugs that surfaced end-to-end:
- send path: encrypt only the freshly appended plaintext tail and post a
single IOWrite wake-up guarded by a bSendPending flag, so back-to-back small
sends (terminal resize -> COMMAND_NEXT -> resize) no longer run concurrent
WSASends over the same mutable buffer (client saw DecryptMessage failures,
with WSAGetLastError() = 6001 / ERROR_DECRYPTION_FAILED);
- client Recv: keep the handshake's unconsumed tail bytes in m_raw and drain
them before the next recv, and log the real SECURITY_STATUS on
DecryptMessage failure instead of the stale Winsock error.
Co-Authored-By: deepseek-v4-pro
Add optional one-way TLS on the device port (6543), enabled per client via a
pinned trust anchor, so new clients encrypt their transport while legacy
plaintext clients keep working unchanged.
Server (server/go): sniff each accepted connection and wrap TLS peers
(ClientHello 0x16 0x03 + handshake type 0x01) in crypto/tls with MinVersion
TLS 1.2; everything else stays on the plaintext path. TLS is enabled only when
both YAMA_TLS_CERT and YAMA_TLS_KEY are set, and a cert/key load failure fails
startup loudly rather than silently serving plaintext.
Client (client/, Windows): Schannel one-way TLS. When the trust anchor is
present — registry value Software\<exehash>\settings\tls_cert (REG_SZ holding
the PEM), with fallback to settings.ini [settings] tls_cert or yama-ca.pem
beside the exe — every outbound connection (main + sub-connections) completes
a TLS 1.2 handshake before the app protocol and is strictly pinned to that
anchor via a chain engine with hExclusiveRoot; any handshake/cert failure is
fail-closed (no plaintext downgrade). Send uses EncryptMessage; Recv uses
DecryptMessage and reads plaintext from the stream-mode DATA segment (bufs[1])
rather than the record header.
Co-Authored-By: deepseek-v4-pro
The file-transfer speed dialog previously sent a separate CMD_FILE_SPEED
(5 bytes) back-to-back before COMMAND_SCREEN_SPY / COMMAND_LIST_DRIVE when
opening remote desktop or file manager. The IOCP send path batches two
consecutive Send2Client calls into one WSASend of the accumulated
OutCompressedBuffer, so the client received the [FILE_SPEED][SCREEN_SPY]
command stream twice - the root cause of two remote-desktop dialogs popping
up on a single click (introduced in commit 76decec).
Fold the two values into MasterSettings (TransferChunkKb / TransferDurationMs,
uint16 each, before Reserved, sizeof unchanged at 1000) so they reach every
client on login and are broadcast to online clients when the dialog saves.
The client persists them into [transfer] on CMD_MASTERSETTING, guarded by
TransferChunkKb != 0 so older masters keep local defaults. CMD_FILE_SPEED(252)
and its client handler remain for backward compatibility, but the master no
longer sends it.
Update the design doc, the three GBK lang files' prompt entry, and the READMEs.
Co-Authored-By: deepseek-v4-pro
Add a connection-management feature to YAMA: a toolbar button (after MCP)
opens a dialog listing the managed host's TCP/UDP connections, with
right-click force-disconnect for IPv4 TCP rows.
Client: a new CConnectionsManager enumerates connections via iphlpapi
(GetExtendedTcpTable/GetExtendedUdpTable, IPv4+IPv6) and answers the
refresh and kill commands over a sub-connection.
Server: a new CConnectionsDlg (DialogBase) shows the list with sortable
columns and a right-click menu; the title-bar icon reuses the toolbar's
connection icon (res/connections.ico).
Fix byte order found during review: iphlpapi returns an IPv4 address as a
network-order DWORD and the port's low 16 bits in network order, so
storing them raw made the dialog show reversed IPv4 addresses and ports.
The client now stores addresses as network-order bytes and ports as host
order (ntohs), and reverses the conversion on disconnect.
Co-Authored-By: deepseek-v4-pro
Add a novice-friendly "File Transfer Speed" dialog under the Advanced menu
that controls file-transfer speed in both directions without touching
SimplePlugins. Speed is passed through InitFileUpload's existing
chunkSizeKb/sendDurationMs arguments, so the precompiled library stays
untouched. Seven presets run from ~0.3 MB/s up to a new 10 MB/s tier plus an
unlimited tier; the chosen value is stored in [transfer] ChunkKb/DurationMs.
Upload reads THIS_CFG at the master's InitFileUpload (2015RemoteDlg.cpp:2225)
and takes effect on restart. Download is pushed to the client over a new
CMD_FILE_SPEED (252) command, sent from OnOnlineFileManager and
OnOnlineDesktopManager before each session opens, and read by all four client
InitFileUpload call sites via GetTransferSpeed so the process-global refcount
cannot latch a stale speed. All Chinese strings go through LangManager, with
en_US/zh_TW/de_DE entries added for the tier names, format strings and hint
text.
Co-Authored-By: deepseek-v4-pro
Add P7 management-control tools behind a single McpManage flag (plus
McpReadonly=0): kill_process, operate_window, control_service and
power_action. The write tools run a ready->confirm lifecycle on their
sub-connection, using a new followUp/followUpSent field to tell the two
arrivals of the same TOKEN_*LIST apart; the write command and the follow-up
refresh command go out as two separate Send2Client calls, with a new
WaitSendIdle() (context.h / Server.h) draining the outbound buffer in between
so the same buffer is never WSASend'd concurrently.
McpSettingsDlg gains a "management operations" checkbox (IDC_MCP_MANAGE), the
main toolbar gains an MCP entry button, and en/de/zh-TW strings are synced.
power_action is fire-and-forget over the main connection (no confirmation).
Verified end-to-end against a real Windows 10/VMware client: kill_process
returns remaining:[] and operate_window/control_service return ok:true with
their confirm packet arriving after an HTTP split. Verification surfaced a
prerequisite parser bug, fixed separately in a72e6eb, where HeaderParser
latched m_bShouldUnmask FALSE on HTTP continuation chunks and dropped the
confirm packet.
Co-Authored-By: deepseek-v4-pro
Implement the P2 upload_file tool to push a local file or directory from the
master to an online Windows host over the existing V2 file-transfer protocol.
The server drives FileBatchTransferWorkerV2 synchronously on the main
connection through a headless callback, reuses the list_files chain for the
overwrite pre-check, and is gated by McpFileTransfer=1 plus McpReadonly=0.
The client main connection never initialized the file-transfer module, so
g_status stayed 0 and RecvFileChunkV2 silently dropped every chunk, truncating
uploads to zero bytes. Add a once-per-process lazy InitFileUpload in the
COMMAND_SEND_FILE_V2 handler that mirrors the FileManager init; the destructor
deliberately does not Uninit so g_status remains 1 across reconnects.
Update the design doc to record the client-side change and correct the
upload_file description to state that sha256 is not returned (V2 has no
receiver-to-sender ACK; integrity is checked client-side and logged only).
Co-Authored-By: deepseek-v4-pro
download_file pulls a remote file or directory back to the controller over
the existing V2 file-transfer protocol. It reuses COMMAND_LIST_DRIVE to open
the file-manager sub-link, then sends CMD_DOWN_FILES_V2 so the Windows client
streams COMMAND_SEND_FILE_V2 chunks and a per-file COMMAND_FILE_COMPLETE_V2
SHA-256 checksum over its own authenticated sub-connection. The server routes
those two packet types into a new per-device FileTransferSession whenever a
download is pending, so the headless path never opens the GUI progress dialog;
the C2C and existing GUI branches are left untouched.
Files are written under a normalized local_dir, and every chunk filename is
resolved and verified to stay inside local_dir (directories included) to block
.. traversal; overwrite=false skips existing files and counts them as skipped.
One transfer per host (mutually exclusive with the one-shot pending registry),
a dedicated McpFileTransfer=0-by-default gate surfaced in the settings dialog,
and audit logging complete the change. upload_file remains future work.
Co-Authored-By: deepseek-v4-pro
Add docs/Mcp_RemoteControl_Design.md, the reference design for an
screenshot-driven AI remote-control feature: the existing get_screenshot for
observation plus remote_open/remote_close/remote_mouse/remote_keyboard for
input injection, reusing COMMAND_SCREEN_PREVIEW_REQ and COMMAND_SCREEN_CONTROL
+ MSG64. Coordinates are normalized (0..1) and mapped server-side to physical
pixels; injection runs over the screen sub-connection opened by
WebService::StartRemoteDesktop's hidden CScreenSpyDlg. Includes a
chat-on-behalf experiment case with clipboard-encoding and window-capture
notes.
Also fold in the get_audit_log encoding correction for
docs/Mcp_Terminal_Design.md, so the two doc changes ship as one commit.
Co-Authored-By: deepseek-v4-pro
Add terminal_open / terminal_exec / terminal_close so an AI can hold one
shell session per Windows host and run a sequence of commands with cwd and
environment preserved, instead of the one-shot exec_command.
The persistent terminal is a separate full-command write capability gated by
McpTerminal (default off) plus McpReadonly=0, with no whitelist and full
audit. One device maps to one terminal session via the shared m_TermSessions
map; idle sessions are swept after 300s. terminal_exec rejects commands that
contain & or | (they corrupt the sentinel control-operator chain) as well as
control characters.
Also harden exec_command and terminal_exec against newline/CR injection, fix a
dangling subCtx after an abrupt shell disconnect, and refresh lastActiveAt on
command completion. Add en/zh-TW translations for the new UI strings and a
design document covering both exec_command and the persistent terminal.
Co-Authored-By: deepseek-v4-pro
Add the two P2d read-only MCP tools, both one-shot sub-links (mode A') on
the existing single-flight pending registry.
list_services sends COMMAND_SERVICES; the client's CServicesManager emits
TOKEN_SERVERLIST on sub-link creation, parsed as 5 null-terminated fields per
record (display_name/service_name/binary_path/status/start_type) with
zero-padding termination. Services are Windows-only, so LNX/MAC hosts get
-32005 up front instead of a 20s timeout.
get_client_log sends COMMAND_QUERY_LOG; the client's CClientLogManager dumps
its full in-memory Logger ring buffer once, then pushes deltas every 3s. The
MCP path takes the first (full) TOKEN_REPORT_LOG and cancels the sub-link so
later deltas stop, while the MFC log dialog keeps receiving deltas on the
non-pending branch. Log text is client ANSI on Windows, decoded by clientType
like process/file names. MessageHandle intercepts both with
IsPending -> TakeMainResponse + CancelIO. Sync the design doc (P2d section +
verification notes; Linux get_client_log timeout is a client-version gap).
Co-Authored-By: deepseek-v4-pro
The get_screenshot tool inherited the MFC thumbnail-preview profile, capping
the frame at 1024px wide (1280 on 4K source), which is too small for AI
vision/OCR of dense UI. Add an optional max_width argument that overrides the
width (clamped to the client's 64..1920 limit) while keeping the RTT-adaptive
jpegQuality. Omitted or 0 keeps the existing thumbnail profile; 1920 yields
near-native resolution (full 1080p on a 1080p source, 1920 wide on 4K). Sync
the design doc.
Co-Authored-By: deepseek-v4-pro
Add the two P2c MCP tools on top of the P2b protocol. list_files lists
drives (COMMAND_LIST_DRIVE -> TOKEN_DRIVE_LIST) or a directory (follow-up
COMMAND_LIST_FILES -> TOKEN_FILE_LIST on the same one-shot sub-link);
get_screenshot reuses the screen-preview RPC with a per-host reqId
correlation so stale or MFC-preview responses fall through to the MFC
path untouched. Both share the P2b single-flight pending registry.
Fix file/process name encoding: the client reads process names, paths and
file names via the ANSI (A) APIs, so they are GBK on Windows regardless of
the CLIENT_CAP_UTF8 capability bit (which governs window titles only).
Decode by clientType (LNX/MAC = UTF-8, Windows = 936) rather than
GetClientEncoding, and convert the list_files path to the client ANSI code
page before sending. Sync Mcp_Phase2_Design.md with the P2c design and the
encoding correction.
Co-Authored-By: deepseek-v4-pro
Add three read-only P2b MCP tools over the existing protocol. list_processes and list_windows trigger the client via COMMAND_SYSTEM / COMMAND_WSLIST on the main connection and receive TOKEN_PSLIST / TOKEN_WSLIST on a one-shot sub-link; get_activity_history uses the main-connection RPC COMMAND_QUERY_ACTIVITY -> TOKEN_REPORT_ACTIVITY. A per-host single-flight pending registry (m_Pending) with a 20s timeout correlates each response to its request and rejects a concurrent request for the same host with -32003. Parsers stop on the first empty record to ignore the client's LocalSize trailing zero padding, and window titles are decoded per the client UTF-8 capability bit. MessageHandle only adds if-guarded branches, so the MFC dialogs are untouched. Sync Mcp_Phase2_Design.md with the mode A'/A architecture, the verification notes, and the registry-backed config location.
Co-Authored-By: deepseek-v4-pro
Add two pure in-memory MCP tools over the online host list. search_hosts filters by name/remark, IP, group and OS (all optional, AND-combined, ASCII case-insensitive substring match); get_host_detail returns one online host by id and answers -32602 for a missing/non-numeric id and -32002 for an unknown or offline id. Sync Mcp_Phase2_Design.md with the second-review corrections and the P2a implementation.
Co-Authored-By: deepseek-v4-pro
Add an optional MCP server (JSON-RPC 2.0 over Streamable HTTP) exposing
an online-host listing tool, protected by a Bearer token. Disabled by
default; configured via a new "Extensions > MCP Settings" dialog.
- McpServer: httplib + JSON-RPC 2.0 dispatch (initialize/ping/tools/list/tools/call)
- McpSettingsDlg: runtime-created dialog for enable/port/bind/token
- HostJson: extract single-host JSON serialization shared with WebService
- FRP: expose MCP port (union with listening/Web ports) when bound to 0.0.0.0
- i18n: en_US / zh_TW translations
Co-Authored-By: deepseek-v4-pro