fix(pi-agents-tmux): the timeline reads the nested toolCall end-event carrier

[Brad (bmethod)]

Follow-up to #1442 for the one finding deferred past its push budget (thread PRRT_kwDORvss6s6Z05wL): the timeline's tool branches now normalize the nested toolCall/tool_call carrier — name, id, status, isError, arguments, result — the shape pinned in pi-session-bridge's event-sanitizer tests. Previously a nested end fell back to name:tool, leaving the paired start marked ✖ no result recorded and emitting a separate neutral tool tool · ok row. One new pin (378 green): a nested end pairs with its plain start and carries the failure tone.

https://claude.ai/code/session_012epxJEzGqT7q3qcFhdZUt5

[linear-code]

VST-327

[macroscopeapp]

Approvability

Verdict: Approved 62168ff

Self-contained bug fix for transcript timeline display formatting. Adds fallback property reads to handle nested event shapes that were previously causing display issues. Changes are limited to display logic with good test coverage.

^{You can customize Macroscope's approvability policy. Learn more.}

[Copilot]

Pull request overview

Note

Copilot was unable to run its full agentic suite in this review.

Adds support for Pi’s nested toolCall/tool_call “carrier” shape so tool executions are correctly paired and failures are reflected in the timeline output.

Changes:

• Extend formatTranscriptForDisplay to derive tool name/id/args/status from nested toolCall objects.
• Propagate nested error/status/result fields when rendering tool_execution_end.
• Add a regression test covering nested toolCall pairing and failure display.

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 4 comments.

File	Description
pi-extensions/pi-agents-tmux/tests/transcript-timeline.test.ts	Adds a test ensuring nested toolCall end events pair with their starts and propagate failure.
pi-extensions/pi-agents-tmux/extensions/subagent/transcript-timeline.ts	Enhances parsing of tool execution events to support nested toolCall fields for pairing and status rendering.

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 1bb30e1187

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "Codex (@codex) review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "Codex (@codex) address that feedback".

Dismissing prior approval to re-evaluate 62168ff

[Copilot]

Pull request overview

Copilot reviewed 4 out of 4 changed files in this pull request and generated no new comments.

Suppressed comments (3)

pi-extensions/pi-agents-tmux/extensions/subagent/transcript-timeline.ts:230

• The nested-carrier extraction logic (call, name, id, plus related fields) is now duplicated across tool_execution_start and tool_execution_end. Consider factoring this into a small helper (e.g., extractToolCallCarrier(event)) that returns { call, name, id, target } (or { name, id } for end) to reduce drift risk as additional shapes/fields are supported.

				// Pi also emits the nested carrier shape { toolCall: { name, id, … } }.
				const call = (event.toolCall ?? event.tool_call) as Record<string, unknown> | undefined;
				const name = stringValue(event.toolName ?? event.tool_name) ?? stringValue(event.name) ?? stringValue(call?.name) ?? "tool";
				const target = primaryToolArgument(event.args ?? event.arguments ?? event.input ?? event.params ?? call?.arguments ?? call?.input);
				const label = target ? `tool ${name} (${oneLine(target, 60)})` : `tool ${name}`;
				const id = stringValue(event.toolCallId ?? event.tool_call_id ?? event.toolUseId ?? event.tool_use_id ?? call?.id) ?? `name:${name.toLowerCase()}`;

pi-extensions/pi-agents-tmux/extensions/subagent/transcript-timeline.ts:247

• The nested-carrier extraction logic (call, name, id, plus related fields) is now duplicated across tool_execution_start and tool_execution_end. Consider factoring this into a small helper (e.g., extractToolCallCarrier(event)) that returns { call, name, id, target } (or { name, id } for end) to reduce drift risk as additional shapes/fields are supported.

				const call = (event.toolCall ?? event.tool_call) as Record<string, unknown> | undefined;
				const name = stringValue(event.toolName ?? event.tool_name) ?? stringValue(event.name) ?? stringValue(call?.name) ?? "tool";
				const id = stringValue(event.toolCallId ?? event.tool_call_id ?? event.toolUseId ?? event.tool_use_id ?? call?.id) ?? `name:${name.toLowerCase()}`;
				const open = openTools.get(id)?.shift();
				const failed = event.isError === true || event.is_error === true || call?.isError === true || call?.is_error === true || stringValue(event.status ?? call?.status) === "error";
				const status = stringValue(event.status ?? call?.status) ?? (failed ? "error" : "ok");
				const resultSize = payloadByteSize(event.result ?? event.output ?? event.content ?? call?.result);

pi-extensions/pi-agents-tmux/CHANGELOG.md:7

• This changelog entry is a single very long line, which makes diffs and release notes harder to read. Consider wrapping it across multiple indented lines (keeping the same bullet) so it stays readable in terminals and typical markdown viewers.

### 2.8.5

- The Transcript timeline reads the nested `toolCall`/`tool_call` end-event carrier — name, id, status, `isError`/`is_error`, arguments, result (the pi-session-bridge event-sanitizer shape). Previously a nested end fell back to `name:tool`, leaving its start marked `✖ no result recorded` beside a separate neutral `tool tool · ok` row, and a nested snake-case error flag rendered as success. Id-less fallback pairing is case-insensitive on the tool name.

[Copilot]

Pull request overview

Copilot reviewed 4 out of 4 changed files in this pull request and generated no new comments.

Suppressed comments (3)

pi-extensions/pi-agents-tmux/extensions/subagent/transcript-timeline.ts:247

• resultSize only checks call?.result for nested carriers, but the top-level code path supports output/content as result payload keys. If the nested carrier uses toolCall.output or toolCall.content (common in some tool result shapes), the timeline will omit the size (and potentially under-report “has result”). Consider expanding the nested lookup to include call?.output/call?.content (and any other supported keys) to match the robustness of the top-level payload handling.

				const failed = event.isError === true || event.is_error === true || call?.isError === true || call?.is_error === true || stringValue(event.status ?? call?.status) === "error";
				const status = stringValue(event.status ?? call?.status) ?? (failed ? "error" : "ok");
				const resultSize = payloadByteSize(event.result ?? event.output ?? event.content ?? call?.result);

pi-extensions/pi-agents-tmux/extensions/subagent/transcript-timeline.ts:230

• The nested-carrier extraction and the name/id derivation logic is duplicated across tool_execution_start and tool_execution_end. To reduce the chance of start/end drifting (e.g., adding a new key on one side but not the other), consider factoring this into a small helper that returns { call, name, id } (and optionally target) used by both cases.

				// Pi also emits the nested carrier shape { toolCall: { name, id, … } }.
				const call = (event.toolCall ?? event.tool_call) as Record<string, unknown> | undefined;
				const name = stringValue(event.toolName ?? event.tool_name) ?? stringValue(event.name) ?? stringValue(call?.name) ?? "tool";
				const target = primaryToolArgument(event.args ?? event.arguments ?? event.input ?? event.params ?? call?.arguments ?? call?.input);
				const label = target ? `tool ${name} (${oneLine(target, 60)})` : `tool ${name}`;
				const id = stringValue(event.toolCallId ?? event.tool_call_id ?? event.toolUseId ?? event.tool_use_id ?? call?.id) ?? `name:${name.toLowerCase()}`;

pi-extensions/pi-agents-tmux/extensions/subagent/transcript-timeline.ts:243

• The nested-carrier extraction and the name/id derivation logic is duplicated across tool_execution_start and tool_execution_end. To reduce the chance of start/end drifting (e.g., adding a new key on one side but not the other), consider factoring this into a small helper that returns { call, name, id } (and optionally target) used by both cases.

				const call = (event.toolCall ?? event.tool_call) as Record<string, unknown> | undefined;
				const name = stringValue(event.toolName ?? event.tool_name) ?? stringValue(event.name) ?? stringValue(call?.name) ?? "tool";
				const id = stringValue(event.toolCallId ?? event.tool_call_id ?? event.toolUseId ?? event.tool_use_id ?? call?.id) ?? `name:${name.toLowerCase()}`;

[github-actions]

Merge queue ejected this PR (removed_from_merge_queue, reason: MANUAL). The auto-merge arm is now DROPPED — nothing will merge this PR until someone re-arms it.

Ejecting merge-group run: https://github.com/vanillagreencom/vstack/actions/runs/32054807496 (Review gate writer)

Failing job(s):

• Post the gate context on the merge-group sha (failure): https://github.com/vanillagreencom/vstack/actions/runs/32054807496/job/95462388696

No usable same-named comparison on the PR head (checks absent, skipped, or still running) — no flake-vs-genuine call is available; inspect the failing run before re-arming.

Automated by merge-queue-ejection-alert (VST-196). This alert never re-arms auto-merge.