Repository navigation
Releases
v0.2.2
Security
- Locked dependencies: dnspython 2.9.2 (CVE-2026-128089) and
opentelemetry-api 1.45.2 (CVE-2026-126360), both transitive.
Changed
- jt-mcp is a WSJT-X server: the package docstring, README and test comments no longer name
JS8Call, and the README no longer points to qso-graph-sim, which is retired.
v0.2.1
Changed
- AGENTS.md and CLAUDE.md: the organization's shared agent guide, and this repository's own.
- README and CHANGELOG name only the programs jt-mcp supports: WSJT-X and JS8Call.
- README: American spelling, as the QSO Graph style guide asks.
- README: style fixes for the QSO Graph style guide (Google Developer Documentation Style Guide, American English).
v0.2.0
Added
- A WSJT-X listener, and the first working tools. WSJT-X's UDP Server
messages, decoded byqso-graph-lib(qso_graph_lib.jt), the code
qso-graph-sim plays them with:jt_cq: who is calling CQ, and what kind (plain, DX, activity such as
POTA or SOTA, contest, region, frequency), the latest per station.jt_status: dial frequency and band, mode, DX call and grid, Tx enabled
or transmitting, your call and grid, per WSJT-X instance.jt_decodes: recent decodes, each read into who is calling whom, grid,
report and sign-off, filtered by callsign, CQ, SNR or band.jt_heard,jt_wspr, andjt_diagnosticsfor the listener itself.
- The listener follows the fleet's "Servers that listen" rule
(QGD-INT-0005/0006): 127.0.0.1 by default, every interface refused,--allow
required on a network address, names refused, a sender not on the list
dropped unread, counted and shown, and everything keyed by what a datagram
says capped. Each rule has a test, and the sender check was removed once to
prove the tests fail without it. - Still read-only: nothing in the package sends a datagram, so WSJT-X's
Reply, Halt Tx and Free Text cannot be sent from here, and the conformance
test asserts it.openWorldHintistruefor the tools that read WSJT-X.
Changed
-
Renamed from
js8call-mcptojt-mcp, and the scope is the JT family.
WSJT-X and JS8Call would have meant one listener each across separate servers.
One server covers them.WSJT-X is first, and it defines the protocol: the
message definitions come from the WSJT-X project,Network/NetworkMessage.hpp
at the schema WSJT-X states, because the project that owns a definition is its
authority. A variant is supported where it works with those definitions; where
it does not, its difference is not implemented here.The GitHub repository redirects from the old name. On PyPI,
js8call-mcp
cannot be renamed: its six releases are yanked (2026-10-29), so nothing
installs it unless pinned, and the project is kept rather than deleted, because
a deleted PyPI name can be registered by anyone and the MCP Registry and old
configurations still name it.jt-mcpis a new project; nothing depended on
the old one beyondget_version_info. Its first release is 0.2.0, because
this repository's history already holdsv0.1.0.
v0.1.0
Changed
-
Python 3.11 is the minimum. 3.10 reached end of life in October 2026
(PEP 619), so it gets no further security
releases on any platform. On Windows it was worse than unsupported:uv's 3.10
is built against OpenSSL 1.1.1w, end of life since September 2023, so HTTPS
from this server ran over an unmaintained TLS library.certifidoes not help
with that — it supplies the root store, not the implementation.3.10 support was already nominal. This package has linted with
ruff target-version = "py311"and type-checked withmypy python_version = "3.11"for some time, so nothing ever verified the 3.10 claim.uv.lockis regenerated and the diff is large, but nothing is upgraded —
it is the 3.10 resolution branch disappearing. Packages that carried two pins,
one per Python branch, now carry the single 3.11+ pin they already resolved to.
An install on 3.11 or later is unchanged.
v0.0.5
Changed
--transport streamable-httpnow says where it bound, and the help text says
it too. It binds127.0.0.1, and FastMCP takes no host option, so it was
never reachable from another machine or from outside a container — and nothing
said so. The bind address is unchanged: loopback-only is deliberate for a
server that holds credentials and has no authentication of its own.
v0.0.4
Fixed
--versionand--helpnow print and exit. They were ignored, so the
command started the MCP server on stdio and waited for JSON-RPC — leaving a
FastMCP banner on screen and an apparently hung terminal. Seven servers in
this fleet already behaved correctly; this brings the rest in line, and a
test asserts it.
v0.0.3
- Move every Node 20 action to its first Node 24 major
- The locked-dependency audit blocked every release
- Four upgrades arrived as 57 pull requests
- Nothing was auditing the lockfile
- Close every open security advisory in the lockfile
- A release is a workflow, not a merge
v0.0.2
MCP specification 2026-07-28 (#5)
- The one tool declares the four annotation hints. Omitting them advertises it as
destructiveHint: true,idempotentHint: false,openWorldHint: true. - There is no
isErrorwork, and the issue says so correctly — "0 occurrences" of error
payloads insrc/, because nothing here can refuse yet. readOnlyHintmeans more here than anywhere else in this fleet. For a reference server it
means "writes no file". HerePLANNEDpromises "No write, no PTT, no transmit control — a
permanent constraint, not a first-version limit", so the hint is a promise about a
transmitter. A hint is a client UX affordance and never a security control, but a false one
is worse than none: a client may skip a confirmation on the strength of it, and what is at the
other end is a radio and an operator's licence.JS-C10asserts it against the code andJS-C11
asserts the promise is still published, so dropping the sentence fails before the rule becomes
unwritten.JS-C10reads identifiers from the parse tree, because prose about PTT is not PTT. The first
version searched lines and flaggedPLANNEDitself — the sentence that promises no PTT, because
it contains the word. Same trap as the error-payload check incq-zones-mcp. It is falsified
both ways: a plantedrig.set_ptt(1)is caught, and a planted sentence mentioning PTT is not.openWorldHint: falseis right today and a judgement call tomorrow.JS-C12fails on the
first socket — which is when somebody must decide, because a running JS8Call is another program
whose decode stream we neither control nor can enumerate, and that is hown1mm-mcpreasoned
about N1MM's UDP broadcast and concludedtrue.tests/test_mcp_conformance.py(JS-C01 onward). Every check falsified — 10 of 10 — with a
control proving the prose case is not a false positive.
Lint (#4)
E501is adopted. One line, thePLANNEDstring, not a defect; it reflows as adjacent
literals so the published value is byte-identical.- The publisher-name exemption the issue was filed for does not apply: this server names no
publisher's field. Zero exemptions, no# noqa: E501.
Not applicable here
- No
certifiwork and no issue for it (qso-graph-devel#73): no outbound call of any kind in
src/.JS-C12namescertifiin its failure message, for the build-out.
v0.0.1
- Placeholder release so the name is held on PyPI while the work is planned
(qso-graph-devel#48). A pending trusted publisher does not reserve a name —
PyPI's documentation is explicit that a name is held only once something is actually
published — so this is a real, installable package rather than an empty repository. - One tool,
get_version_info, reporting the development status, what is planned, and
where the work is tracked. The server's instructions say the same, so a client listing
it sees the truth before calling anything.