PR #3648 review (codex): the first iteration emitted every envelope
path alongside one combined digest, so a multi-file payload that added
`: any` to a README.md hunk and merely touched src/ok.ts would surface
a *.ts path in the TTSR match context and trip the bundled
tool:edit(*.ts) ts-no-any rule on text that belonged to the Markdown
hunk — aborting valid edits under interruptMode:always.
Add a per-file matcherEntries(args) hook on AgentTool / EditStreaming-
Strategy returning [{ path, digest }] entries, one per touched file
(same-path sections/hunks merged):
- replace / patch: one entry from the top-level path + matcherDigest
- hashline: regex-split by [path#TAG] section, body added-lines per
entry (tolerant of streaming partial payloads)
- apply_patch: expandApplyPatchToPreviewEntries grouped by path
AgentSession.#checkTtsrStream / #checkTtsrAstStream now prefer
matcherEntries and iterate per-file with isolated filePaths + streamKey,
so each file's buffer and repeat-tracking are independent. Tools
without matcherEntries keep the existing combined matcherDigest +
matcherPaths path.
- Standardized retry logic to ensure mocked assistant turns and generic aborts share the same reset path as live provider failures.
- Introduced a classification helper that forces retry policy alignment with the currently active session model.
- Enabled retries for generic reason-less abort sentinels and stale OpenAI response errors.
AgentSession's TTSR match context only scanned top-level path/paths
arguments, so hashline and apply_patch edit streams (whose only target
path lives inside the wire payload — section headers or envelope
markers — not as a top-level argument) arrived without any filePaths
and silently skipped path-scoped rules like the bundled ts-no-any
(scope: tool:edit(*.ts)).
Add an optional AgentTool.matcherPaths(args) hook, companion to the
existing matcherDigest(args), so tools whose wire grammar embeds paths
can surface them. Implement on each edit streaming strategy:
- replace / patch: top-level path
- hashline: parse [path#TAG] (and tag-less [path]) section headers
tolerant of streaming partial payloads
- apply_patch: parse *** Add/Update/Delete File: markers, also tolerant
of pre-End-Patch buffers
AgentSession.#getTtsrToolMatchContext consults tool.matcherPaths first,
normalising its output through the existing path-candidate helper, and
falls back to the generic top-level argument scan for tools that don't
implement it.
Fixes#3646
- Replaced functional streaming logic with a stateful `CodexStreamProcessor` class.
- Converted `CodexStreamRuntime` into a class with explicit state initializers and lifecycle methods.
- Encapsulated tool call delta tracking and interruption logic within dedicated class methods.
- Integrated `textSignature` generation into text-type output blocks using versioned encoders.
AgentSession.#buildAdvisorRuntime constructed the advisor Agent without the provider-shaping options the SDK installs on the main agent: the streamFn wrapper that applies providers.openrouterVariant / providers.antigravityEndpoint / providers.maxInFlightRequests / model.loopGuard.*, the onPayload/onResponse/onSseEvent hooks, the shared providerSessionState map, transformProviderContext (snapcompact, secret obfuscation, image clamping), and a stable promptCacheKey. Advisor turns therefore dropped the OpenRouter sticky-routing variant suffix, used a different prompt_cache_key than the main turn, and skipped the per-session provider hooks — producing intermittent OpenRouter response-cache misses across consecutive advisor calls.
Extract the inline streamFn wrapper in sdk.ts into a shared createSettingsAwareStreamFn helper (packages/coding-agent/src/session/settings-stream-fn.ts) and pass it (plus transformProviderContext) through AgentSessionConfig as advisorStreamFn / transformProviderContext. #buildAdvisorRuntime now hands the advisor Agent the same streamFn, hooks, providerSessionState, promptCacheKey (= advisor session id), and transformProviderContext as the main turn. Adds getAdvisorAgent() accessor on AgentSession for diagnostics and parity tests.
Fixes#3639
- Removed the pi dialect implementation and associated source files.
- Updated dialect resolution, factory registration, and type definitions to exclude pi.
- Cleaned up settings schema and user options to remove pi-related configurations.
- Deleted corresponding test suites covering pi dialect functionality, in-band tools, and examples.
- Removed "running" status and hub hint details from the subagent badge text.
- Updated relevant status line tests to expect the simplified badge format.
- Updated `reconcileTitleCasing` to ignore purely uppercase words when restoring source casing.
- Restricted restoration logic to mixed-case identifiers (e.g., `TinyVMM`, `iOS`) to avoid overriding the model's clean sentence case with emphatic user shouting.
- Added regression tests to ensure all-caps input does not trigger title re-shouting.
- Added `tiny` as a first-class model role to override online models for lightweight background tasks.
- Updated session title generation, auto-thinking difficulty classification, unexpected-stop detection, and mnemopi backend to resolve via the `tiny` role before falling back to `smol`.
- Updated configuration schema and documentation to reflect the new role precedence.
On a long subagent-heavy run the TUI thought stream froze for 1.5–4 s
at a time, with the loop watchdog blaming `subagent:*` phases and the
log carrying ~1,255 `Skipping mid-run compaction because turn persistence
is out of order` debug lines per session (issue #3629).
`#persistTurnMessagesForMidRunCompaction` and the persistence-check
helpers it calls were doing two O(n²) things per `onTurnEnd`:
1. Every call rebuilt the branch path via `SessionManager.getBranch()`
(O(branchSize) `unshift` per call), once per turn message twice (once
in `#sessionMessageAlreadyPersisted`, once in
`#hasPersistedLaterTurnMessage`).
2. Every pairwise message comparison serialized full message content
through `#messageValueSignature` (=`JSON.stringify`), even though
the content was almost always orthogonal to the identity decision.
Replace the structural comparator with a stable persistence key
(timestamp + role-specific discriminators) extracted to a sibling module
`session/turn-persistence.ts` and drive the planner off one snapshot of
the branch built per `onTurnEnd`. Content equality is retained as the
slow-path tiebreaker for the rare case where two messages collide on the
cheap key (e.g. two assistant turns at the same millisecond with
`undefined` responseId — the shape the test harness emits).
The new helpers (`sessionMessagePersistenceKey`, `planTurnPersistence`,
`sameMessageContent`) are pure functions covered by direct unit tests
in `test/turn-persistence.test.ts`; the existing mid-run / eager
compaction integration tests prove the end-to-end behavior is preserved.
Fixes#3629
Filtered marketplace-managed runtime symlinks out of the npm plugin listing and OMP extension-package status provider while keeping marketplace installs available to the runtime loader. Added regressions for both duplicate surfaces.
Fixes#3628
- Added formatAdvisorContextPrompt to render project context files into the advisor's system prompt.
- Updated AgentSession to accept and inject advisorContextPrompt into the session system prompt.
- Registered project context files for the advisor to ensure the reviewer evaluates the agent against standing project instructions like AGENTS.md.
Resolved live ACP generate_image payloads through the blob store before emitting image content while keeping rawOutput compact.
Added regression coverage for content[] image blocks and details.images entries without duplicating blob refs as fallback text.
Fixes#3623
Expanded environment variable placeholders in Claude marketplace plugin MCP url and headers before registration. Added a regression test covering context7-style HTTP server headers.\n\nFixes #3621
Initialized the Z.AI Streamable HTTP MCP session before calling web_search_prime and preserved the returned session id on subsequent requests.
Added regression coverage for the authenticated MCP request sequence.
Fixes#3619
- Redesigned the Todo HUD as a connector tree with fixed-budget stage previews.
- Anchored status and HUD containers to prevent redundant UI elements in terminal scrollback.
- Implemented tree-based rendering for project phases and tasks while removing dynamic border rules.
- Upgraded `sherpa-onnx` and related packages to support current infrastructure.
Address PR #3602 review feedback from chatgpt-codex-connector:
when a stdin read carries the empty bracketed paste followed by
a trailing keystroke (a user pressing Enter right after Cmd+V),
the pre-fix paste path was fire-and-forget. The trailing byte
processed synchronously while the clipboard image read was still
pending, so submit ran against an empty pendingImages and the
image landed on the next draft instead.
CustomEditor now tracks in-flight pastes with #pasteInFlight and
buffers subsequent input into #pendingInput. #trackAsyncPaste
increments the counter, awaits the paste promise, decrements, and
drains the queue through handleInput (so requeueing still works
if a drained chunk triggers another async paste).
For an assembled paste whose remaining bytes are present in the
same call, those bytes are pushed onto #pendingInput before the
async paste starts, so they always run AFTER it settles. The
text-paste branch stays sync and drains its own queue inline.
New repro test asserts the call ordering: paste:start fires, the
queued Enter does NOT, and only after the paste promise settles
does Enter dispatch.
- Enhanced the edit renderer to support visual tracking of delete and move/rename operations.
- Updated diff computation to correctly handle file-level changes and suppress erroneous "No changes" warnings.
- Improved terminal output with a clearer activity indicator and accurate state representation during multi-file operations.
- Extended the rendering pipeline to display source-to-destination paths for file renames and added validation tests for edit workflows.
- Added `rewrite-changelog.ts` and `fix-changelogs.ts` utilities to automate the consolidation of release notes using LLM-assisted processing.
- Updated multiple internal changelog files by consolidating redundant entries and improving phrasing for readability.
- Implemented `previewLine` utility in `coding-agent` to prevent visual spillover in status rows by managing text truncation and whitespace.
- Updated `package.json` with new workflow scripts for managing package-level change histories and documentation indexes.
Address PR #3602 review feedback from chatgpt-codex-connector:
when the terminal fragments a bracketed paste across stdin chunks
(\x1b[200~ in one read, \x1b[201~ in the next — Windows Terminal
under load, certain SSH muxes, tmux extended-keys passthrough),
the previous single-chunk `isEmptyBracketedPaste` /
`extractBracketedImagePastePaths` guards never saw both markers
in the same `handleInput` call. The inherited
`BracketedPasteHandler` then buffered the run as a zero-length
text paste and Cmd+V still disappeared.
CustomEditor now owns its own BracketedPasteHandler that runs
ahead of `super.handleInput`, so split bracketed pastes resolve
to a single assembled payload before any routing decision:
- empty payload -> onPasteImage (Cmd+V macOS image-only screenshot)
- image-file paths -> onPasteImagePath (#3506 also gains split-chunk
coverage as a bonus)
- everything else -> base editor's public `pasteText` so the
`[Paste #N]` markers, autocomplete, and undo state stay intact
Removed the now-redundant single-chunk `isEmptyBracketedPaste`
helper. New repro tests cover the split-chunk empty paste, the
split-chunk image-file path, and a split-chunk text paste
forwarding exactly once to the base editor.
macOS terminals (iTerm2, Terminal.app, Warp, Ghostty without OSC 5522,
…) intercept Cmd+V and read NSPasteboardTypeString first. For an
image-only clipboard (Cmd+Shift+5 screenshot saved to clipboard, Chrome
image copy, …) that read returns empty, so the terminal forwards a
complete-but-empty bracketed paste — '\x1b[200~\x1b[201~' — to the app.
CustomEditor.handleInput inserted that empty payload and the keystroke
disappeared, forcing users back to Ctrl+V (which is never intercepted
and already routes through handleImagePaste).
Add isEmptyBracketedPaste and route a complete, empty-or-whitespace-only
bracketed paste through the same onPasteImage smart reader the
app.clipboard.pasteImage keybind uses, so Cmd+V attaches the clipboard
image (or falls back to the #1628 smart text paste / 'clipboard is
empty' diagnostic) instead of silent nothing. Bracketed pastes carrying
any text (including the explicit image-file path branch from #3506)
keep their existing routing.
Fixes#3601
can1357