Repository navigation
Conversation
Story 66 wants an Advanced section on the Director tab that points any MCP-capable harness at ai-buddy. This is the registration pass behind it: what the generated snippet contains for each of the five harnesses, and where one box is not enough. It also names the blocker. ai-buddy has no surface that reveals the token, so none of these procedures can be completed by a user today. Five options, with what each costs against ADR-0023 and ADR-0026. Section 6 of this pass shipped separately in #743; the section here points at it rather than repeating it. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
|
Closing this as superseded. The owner was right that it is outdated, and I should have checked before opening it. The document's central claim is false on current main. It says "None of the above is reachable today, because ai-buddy does not tell anyone the token." #577 surfaced it: Worse, the findings it presents as new are already in the shipped code as doc comments on
#580 checked all five snippets against the installed CLIs, and So every section is merged, implemented, or moot: §"Prompting is not a route" shipped in #743, §"What the Settings box has to contain" and §"Where one box is not enough" are My verification error, for the record. I did check this branch against main before publishing, but I checked the wrong thing: that Nothing here is worth salvaging into a separate document. The branch is deleted. — Claude Code agent |
Rescues the registration pass that has been sitting unpushed on
research/byo-harness-mcpsince before #743. One new file,
docs/research/byo-harness-mcp-registration.md. No code.#743 carried one of this pass's seven sections into
byo-harness-mcp-reload.md§6. Theother six have never been published, and they are the half that bears on the Settings work.
What it answers
Story 66 wants an Advanced section on the Director tab: a harness picker, a generated
snippet in a copyable box, per-harness instructions under it. This says what that box has
to contain for each of the five harnesses, and where one box is not enough.
All five take the loopback HTTP endpoint with
Authorization: Bearer <token>as anordinary configured header, so ADR-0026's stdio shim is a fallback here, not a
necessity. Four were driven against a replica of
mcp_http.rs's gate and connected. Nonesends an
Originheader, so the DNS-rebinding refusal never fires.Concrete things the snippet generator needs to know:
Bearerprefix. Three of five templates add itthemselves; a pre-prefixed token yields
Bearer Bearer …in the other two.name prints "already exists", exits 0, and leaves the stale URL and token in place. Both
values change every app run, so the
removeis what makes the snippet re-runnable.-s projector grok's--scope project. Both write a committablefile, which is the only failure mode here that puts the token somewhere the user did not
expect.
claude mcp list,hermes mcp test ai-buddy,opencode mcp list,grok mcp doctor ai-buddy. Four wererun; they report connected/failed and a tool count without starting a session.
is somewhere the user returns to, not a setup step. Only hermes makes it a paste.
The blocker it names
None of it is reachable today.
Endpointinsrc-tauri/src/mcp_http.rsderivesCloneand nothing else — noDebug, noSerialize— and the token leaves it throughexactly two methods, both called only from
harness.rsto compose themcpServersentryfor a Harness ai-buddy spawned itself. No Tauri command, no settings row, no CLI flag
prints it. That is ADR-0023 and ADR-0026 working as intended for the spawned case, and it
is what story 66 cannot live with.
A second process cannot solve it:
ai-buddy --print-mcp-endpointwould have a differenttoken, and making it ask the running app means an unauthenticated loopback route handing
out the credential. The reveal has to come from the running app's own UI.
Five options are laid out with their costs. The document's own reading is 1 + 2 —
reveal it in the generated snippet, and refuse the shared-file scopes — with 4 (fixed
port, per-run token) held in reserve for when the re-paste is what people actually
complain about. 3 (a stable token on disk) reverses ADR-0026 on its own terms and
should arrive as its own ADR, not as an implementation detail of a settings pane.
This is research, not a decision. It does not amend any ADR.
Evidence discipline
Every claim is marked ran or read, and the two are never mixed in one claim. The
What is not verifiedsection is honest about the soft parts: codex is entirelyunverified (not installed here), nothing was tested against the running binary, no
harness was observed actually calling a tool, and the mid-session column is the weakest
row in the table. A borrowed machine with codex settles the largest gap in a minute.
Verification
Probes ran at
d55beb74. Before publishing I re-checked every claim this document makesabout ai-buddy's own source against
772026cf:Endpointstill derivesClonealone,and
authorization()andstdio_env()are still its only two ways out. The preamble saysso.
node --test tests/passes (203) andpre-commitis clean on the file.— Claude Code agent
🤖 Generated with Claude Code