fix: stop a run on an auth failure instead of retrying it (#86)

CLAUDE.md

@@ -20,7 +20,7 @@ docker compose up -d
 docker compose logs -f claude-chat
 ```
 
-No linting and no build step configured. `npm test` chains 76 test files under `test/`: 19 DOM-less render/UI-logic tests (`test/render/*.test.mjs`, run through `node --test`) plus 57 plain-`node` suites in `test/` covering the overload detector, env load order, multi-agent results, terminals, bots, telegram, updates, kanban scheduling, the Kanban card's run-settings badges (`kanban-run-badges.test.js` pins the card's model/effort/engine chain against the one `startTask` actually resolves — the two live in different files and the card silently lies when they drift), the board-only `create_task` status (`task-backlog.test.js`), i18n completeness, the config precedence resolver plus its secret masking, usage-limit detection, the filesystem path guard (including the SVG sandbox header and the symlink rule on the `@`-mention search endpoints) plus the tunnel-blocks-terminal rule, WS session re-subscription, the SSH remote CLI-session import, the live engine pane / interactive-prompt watchdog, the cross-project global workspace aggregation, the rule that an SSH credential never leaves the server process, the Windows command-quoting oracle, the auth token lifecycle, the multi-agent dependency scheduler (waves, plan sanitising, and the rule that a failure warning must survive dep-context truncation), the SSH stream parser's three guards, the SSH run's termination guarantee (`ssh-termination.test.js` drives `ClaudeSSH.send()` through a fake ssh2 `Client` in `require.cache` and asserts `onDone` fires EXACTLY once on every ending — a missed one hangs the chat forever, a doubled one re-emits stderr) and the recovery contract of "Restart Session" (`session-restart.test.js` boots a real server against a fake `claude` that never exits, then pins that a restart ABORTS that turn and releases the session instead of refusing), the remote non-interactive shell environment (`remote-env.test.js`, which runs the generated prelude through real `bash -lc`: it must parse, print nothing on stdout, and never end on a false test — the caller chains `&& claude …` behind it), the remote CLI-list framing parser, the bot inbox's SQL seam (`bot-inbox.test.js` pins that `from_bot AS "from"` keeps the exact key `planInboxDelivery` reads — rename one without the other and every letter is silently retired as malformed), the one-time config/.env migration onto CCS_CONFIG_PATH and the mid-task clarification delivery contract on the subscription engine (`interrupt-delivery.test.js` — pins that the tmux injection block sits BEFORE the poll loop's completion `break`, that draining does not imply delivery, that a failed paste is re-queued and warns non-terminally, and that the task runner passes the same callbacks the chat path does), the CLAUDE.md / AGENTS.md discovery rules (`agents-md.test.js`, which also pins that AGENTS.md reaches the subprocess as `--append-system-prompt` and never as `--system-prompt`), and the remote file browser's three guard layers (`remote-files.test.js` runs the generated POSIX script through a real `/bin/sh` against a temp tree that contains symlinks OUT of the project; `remote-files-api.test.js` boots a server against a fake remote via `CCS_REMOTE_EXEC_HOOK` and drives `/api/files` the way the SPA does), the editor deep links (`editor-links.test.js` pins the two URI shapes literally — the browser link puts `vscode-remote` in the AUTHORITY and the CLI argument puts it in the SCHEME, and collapsing the two silently breaks one path; `editor-open-api.test.js` boots a real server with `PATH` pointed at an EMPTY directory, which both makes the `opened:'client'` fallback deterministic and guarantees the suite never launches an editor window on a developer's desktop), and the new-chat defaults chain (`chat-defaults.test.js` pins the pure resolver — the built-ins are asserted to be exactly what the SPA hardcoded before #58, and the choice lists to be exactly the toolbar's `data-v` sets and `MODEL_MAP`'s aliases; `chat-defaults-api.test.js` boots a real server in a throwaway `APP_DIR` and pins that a project writes back a SPARSE override object — a five-key snapshot passes every other assertion in that file and still breaks the feature). On the render side, `tables.test.mjs` also pins the ReDoS bound in renderMd step 3.4, `xss.test.mjs` runs 24 adversarial payloads end-to-end, and `forged-tokens.test.mjs` covers the case where user text contains the renderer's own placeholder control bytes, and `pane-font.test.mjs` pins the clamp DIRECTION of `_fitEnginePaneFont` (a wide engine pane may only shrink; a narrow split pane must be allowed to grow). `script-scope.test.mjs` pins which `<script>` block a helper is declared in — declarations hoist only within their own block, so a helper used by `loadSess()` must not live in the terminal block at the bottom of the file. Note the glob: a file under `test/render/` whose name does not end in `.test.mjs` is NEVER run — `_load.selftest.mjs` sat there unexecuted until it was renamed to `loader.test.mjs`. It runs serially and aborts on the first failing file. `.github/workflows/ci.yml` runs it on every push and PR to `main` (tmux installed, so the five tmux-dependent suites do not self-skip).
+No linting and no build step configured. `npm test` chains 77 test files under `test/`: 19 DOM-less render/UI-logic tests (`test/render/*.test.mjs`, run through `node --test`) plus 58 plain-`node` suites in `test/` covering the overload detector, env load order, multi-agent results, terminals, bots, telegram, updates, kanban scheduling, the Kanban card's run-settings badges (`kanban-run-badges.test.js` pins the card's model/effort/engine chain against the one `startTask` actually resolves — the two live in different files and the card silently lies when they drift), the board-only `create_task` status (`task-backlog.test.js`), i18n completeness, the config precedence resolver plus its secret masking, usage-limit detection, authentication-failure classification (`auth-errors.test.js`, which pins that the detector runs BEFORE the auto-continue in all three agent loops — a reorder there silently restores #86), the filesystem path guard (including the SVG sandbox header and the symlink rule on the `@`-mention search endpoints) plus the tunnel-blocks-terminal rule, WS session re-subscription, the SSH remote CLI-session import, the live engine pane / interactive-prompt watchdog, the cross-project global workspace aggregation, the rule that an SSH credential never leaves the server process, the Windows command-quoting oracle, the auth token lifecycle, the multi-agent dependency scheduler (waves, plan sanitising, and the rule that a failure warning must survive dep-context truncation), the SSH stream parser's three guards, the SSH run's termination guarantee (`ssh-termination.test.js` drives `ClaudeSSH.send()` through a fake ssh2 `Client` in `require.cache` and asserts `onDone` fires EXACTLY once on every ending — a missed one hangs the chat forever, a doubled one re-emits stderr) and the recovery contract of "Restart Session" (`session-restart.test.js` boots a real server against a fake `claude` that never exits, then pins that a restart ABORTS that turn and releases the session instead of refusing), the remote non-interactive shell environment (`remote-env.test.js`, which runs the generated prelude through real `bash -lc`: it must parse, print nothing on stdout, and never end on a false test — the caller chains `&& claude …` behind it), the remote CLI-list framing parser, the bot inbox's SQL seam (`bot-inbox.test.js` pins that `from_bot AS "from"` keeps the exact key `planInboxDelivery` reads — rename one without the other and every letter is silently retired as malformed), the one-time config/.env migration onto CCS_CONFIG_PATH and the mid-task clarification delivery contract on the subscription engine (`interrupt-delivery.test.js` — pins that the tmux injection block sits BEFORE the poll loop's completion `break`, that draining does not imply delivery, that a failed paste is re-queued and warns non-terminally, and that the task runner passes the same callbacks the chat path does), the CLAUDE.md / AGENTS.md discovery rules (`agents-md.test.js`, which also pins that AGENTS.md reaches the subprocess as `--append-system-prompt` and never as `--system-prompt`), and the remote file browser's three guard layers (`remote-files.test.js` runs the generated POSIX script through a real `/bin/sh` against a temp tree that contains symlinks OUT of the project; `remote-files-api.test.js` boots a server against a fake remote via `CCS_REMOTE_EXEC_HOOK` and drives `/api/files` the way the SPA does), the editor deep links (`editor-links.test.js` pins the two URI shapes literally — the browser link puts `vscode-remote` in the AUTHORITY and the CLI argument puts it in the SCHEME, and collapsing the two silently breaks one path; `editor-open-api.test.js` boots a real server with `PATH` pointed at an EMPTY directory, which both makes the `opened:'client'` fallback deterministic and guarantees the suite never launches an editor window on a developer's desktop), and the new-chat defaults chain (`chat-defaults.test.js` pins the pure resolver — the built-ins are asserted to be exactly what the SPA hardcoded before #58, and the choice lists to be exactly the toolbar's `data-v` sets and `MODEL_MAP`'s aliases; `chat-defaults-api.test.js` boots a real server in a throwaway `APP_DIR` and pins that a project writes back a SPARSE override object — a five-key snapshot passes every other assertion in that file and still breaks the feature). On the render side, `tables.test.mjs` also pins the ReDoS bound in renderMd step 3.4, `xss.test.mjs` runs 24 adversarial payloads end-to-end, and `forged-tokens.test.mjs` covers the case where user text contains the renderer's own placeholder control bytes, and `pane-font.test.mjs` pins the clamp DIRECTION of `_fitEnginePaneFont` (a wide engine pane may only shrink; a narrow split pane must be allowed to grow). `script-scope.test.mjs` pins which `<script>` block a helper is declared in — declarations hoist only within their own block, so a helper used by `loadSess()` must not live in the terminal block at the bottom of the file. Note the glob: a file under `test/render/` whose name does not end in `.test.mjs` is NEVER run — `_load.selftest.mjs` sat there unexecuted until it was renamed to `loader.test.mjs`. It runs serially and aborts on the first failing file. `.github/workflows/ci.yml` runs it on every push and PR to `main` (tmux installed, so the five tmux-dependent suites do not self-skip).
 
 ## Architecture
 
@@ -793,6 +793,49 @@ would be noise on every long chat. It is applied to the FIRST retry notice as we
 exhausted one — `hit the 50-turn limit (used 3)` contradicts itself, and a user who stops
 reading after the first retry never reaches the corrected sentence.
 
+### A stop that retrying cannot fix (issue #86, auth-errors.js)
+
+`"Failed to authenticate: OAuth session expired and could not be refreshed"` is the
+THIRD way a turn stops through no fault of the agent, and the retry machinery handled
+it worst. The two in `rate-limit-utils.js` both clear on their own — an overload after a
+backoff, a quota at `resetsAt`. An auth failure never does; a human must re-authenticate
+the CLI. But it is not `subtype:'success'`, so all three agent loops AUTO-CONTINUED it:
+each retry failed instantly with the identical error, the whole `MAX_AUTO_CONTINUES`
+budget emptied in seconds, and a chain task then retried twice more — ending as a generic
+`agent_incomplete` that named nothing. The report was "work stops and I am not told why";
+the burnt budget was the part nobody could see.
+
+- **This module refreshes nothing, and cannot.** The OAuth tokens belong to the `claude`
+  CLI's own credentials store; the refresh exchange needs a refresh token and client
+  credentials this server has never held and must not hold. "Refresh automatically" is
+  the CLI's job — when it reports that the refresh FAILED, the only correct move here is
+  to stop at once and name the action. Detecting it precisely and REFUSING to retry is
+  the fix, not a lesser version of one.
+- **The clean-success guard is what keeps it usable in this repo.** These are English
+  CLI strings, and issue #86, `auth-errors.js` and its own test all contain the phrase
+  "OAuth session expired" — an agent asked to read any of them reproduces it in a turn
+  that SUCCEEDED. `detectAuthError()` returns null for `subtype:'success'` with no error
+  flag, exactly as `shouldRetryOverload()` does. A turn that ended well was not blocked
+  by authentication, whatever its text says.
+- **Anchors are CLI/API-internal wording, never keywords.** A bare `/login` or
+  `unauthorized` is excluded deliberately: this project serves its own `/login` route and
+  answers `{error:'unauthorized'}` from its own middleware, so either would fire on a run
+  that merely reads `server.js`. The set is also DISJOINT from `rate-limit-utils.js` in
+  both directions — reading a quota banner as an auth stop would strand work that would
+  have resumed by itself.
+- **The ordering is the fix, so the test pins indices.** `auth-errors.test.js` compares
+  the position of the detector against the success break and `taskContinueCount++` in
+  `server.js`. Moving the block below the auto-continue restores the bug while every
+  behavioural assertion still passes — verified by mutation.
+- **The card has to say it.** `failure_reason` carries `auth_error:<kind> <line>`, mirroring
+  the parseable `usage_limit:<unix> …` prefix, and `kbAuthStop()` renders a 🔐 badge.
+  Without it an auth stop is indistinguishable from an ordinary failure on the board,
+  which is the "nothing tells me why" half of the report.
+- **`claudeCliStatus().authenticated` does NOT cover this, by construction.** It tests
+  that `~/.claude/.credentials.json` EXISTS — an expired token's file exists too. Telling
+  a live token from a dead one needs a network round trip, so the preflight flag stays a
+  cheap "is anything configured" check and the stop is caught reactively, mid-turn.
+
 ### Composer geometry and terminal-pane control
 
 Four reports that all reduce to "the UI moved or stopped listening". Pinned by
@@ -914,7 +957,7 @@ message replays the chat's history into a fresh Claude session rather than losin
 
 ## How to Verify Changes
 
-`npm test` runs 76 test files under `test/` (19 `test/render/*.test.mjs` + 57 `test/*.test.js`), and `.github/workflows/ci.yml` runs the same command on every push and PR to `main`. Nothing covers the live browser/WebSocket path, so also verify that manually:
+`npm test` runs 77 test files under `test/` (19 `test/render/*.test.mjs` + 58 `test/*.test.js`), and `.github/workflows/ci.yml` runs the same command on every push and PR to `main`. Nothing covers the live browser/WebSocket path, so also verify that manually:
 
 ```bash
 # 1. Start server