@capagents/forgeqe 1.0.5 → 1.0.7

package/HOW_TO_USE.md

@@ -66,6 +66,7 @@ Everything ForgeQE can do, by area. ForgeQE's agent answers "can you…?" questi
 | Release | **Requirement coverage and release readiness**: Gives requirement coverage and a GO / GO WITH RISKS / NO-GO verdict for a sprint, area, epic, or stories, from test cases, results, bugs, and flaky tests; can publish it to the wiki. | `release_readiness` | [docs](#requirement-coverage-and-release-readiness) |
 | Operations | **Scheduled monitoring with alerts**: Runs checks on a schedule and alerts Teams or email when one keeps failing, and again when it recovers. | `monitor` | [docs](#9-monitoring) |
 | Operations | **Teams and email notifications**: Tells the team in Microsoft Teams channels or by email when a suite fails or passes, when a release readiness verdict is in, and when the domain evaluation falls below its gate, with the failing cases, their bugs, and links to the report and the CI build. Each channel picks its events; the agent also posts a summary to a channel when asked. | `notify` | [docs](#17-notifications) |
+| Release | **Reports you can share**: Writes up what ForgeQE's agent found, such as release status across projects, open defects, or pipeline health, as a Markdown file and an HTML page in the reports folder, with a summary in the chat and links to open them. | `write_report` | [docs](#reports-from-the-agent) |
 | Operations | **CI pipelines**: Writes a GitHub Actions or Azure Pipelines workflow that runs ForgeQE, stores the secrets it needs, pushes a branch, and makes the pipeline green. | `push_branch`, `set_ci_secret` | [docs](#8-pipelines) |
 | Systems | **Work with Azure DevOps, Jira, and Confluence**: Reads anything in the configured systems (work items, issues, queries, boards, pages), makes changes the user asks for, and builds Azure DevOps dashboards and wiki pages. | `query_api`, `read_confluence`, `create_ado_dashboard`, `write_ado_wiki_page` | [docs](README.md#azure-devops-jira-and-other-systems) |
 | Setup | **Setup, skills, and documentation**: Shows the model profiles and MCP servers, installs or repairs the browser engine, creates a forgeqe.yaml, manages skills, and answers questions about ForgeQE from its documentation and this list. | `forgeqe_setup`, `capabilities` | [docs](#2-scaffold-config) |
@@ -366,6 +367,18 @@ While a run is going, the bar under the prompt shows a spinner, the tokens used,
 
 The text is the agent's own result. Failed runs show `✗ Task failed` with the reason, and stopped runs show `■ Run stopped`. `forgeqe run` prints the same summary at the end.
 
+### Reports from the agent
+
+Ask for a report and ForgeQE's agent writes one you can share: "create a release report for all projects", "write up the open defects", "summarise this sprint's test results as a report". It gathers the facts with its tools (Azure DevOps, Jira, the runs it made), then saves the report with `write_report` as a Markdown file and an HTML page in the reports folder (`export.out_dir`, `.forgeqe/reports/` by default):
+
+```
+.forgeqe/reports/release-status-all-projects-20261009-001512/
+  report.md     # to edit, paste into a wiki, or attach
+  report.html   # to open or share; light and dark, tables scroll on a phone
+```
+
+The chat gets a short summary and a card with both paths as links, and **Open the report** appears as a next step. The report holds only what the agent read in that request. To post it, ask for it ("post the report to the release channel") once `notify.channels` is set up; the reports folder is linked when it is served (`notify.report_url`).
+
 ### The screen
 
 On a terminal 120 columns wide or more, the shell splits in two once you send a request:
@@ -375,6 +388,12 @@ On a terminal 120 columns wide or more, the shell splits in two once you send a
 
 On a narrower terminal the chat fills the screen, and Ctrl-O switches between the chat and the same panels.
 
+**Buttons.** Under the prompt: **✦ Refine** (Ctrl-R; it fills in once there is text to refine) and **⌕ Search** (Ctrl-P, the command palette); while a request runs, **■ Stop** (Esc) takes Refine's place. In the status line above the panels: **▦ Panels** (Shift-Tab) and **⤢ Wide chat**, which gives the conversation the whole width for reading a long answer; **◧ Split view**, under the prompt, brings the panels back. On a narrow terminal, **◧ Details** (Ctrl-O) and **⇄ Profile** (Tab) are under the prompt. Each is clickable.
+
+**Progress.** While a request runs, the status line above the panels shows a bar for how far the agent's plan has got (`███████░░░░░ 2/3`), or the suite's cases while a suite runs.
+
+**Replies.** The agent's answers lead with the answer, then short headed sections; lists keep their bullets in a column, so a wrapped line stays under its text instead of running back to the margin.
+
 **Moving between panels.** Shift-Tab (or a click) moves the keys from the prompt to Plan, Activity, then Files; the focused panel's border lights up and shows its keys. Esc, or typing, goes back to the prompt.
 
 | Panel | Keys |
@@ -393,7 +412,7 @@ On a narrower terminal the chat fills the screen, and Ctrl-O switches between th
 
 **Refine.** Ctrl-R (or *Refine my request* in the palette) turns what you typed into a precise request before it is sent. A read-only agent checks it against this setup: what ForgeQE can do here, the systems that are set up, the project's sprints and test cases, and recent runs. It shows the refined request with what it **checked** ✓, what it **assumed** ~, and what is **missing** ?. Enter uses it, `e` puts it in the prompt to edit, Esc keeps yours. Refine never runs anything, writes files, or changes a work item.
 
-**Next steps.** When a request finishes, up to three suggestions appear above the prompt, such as *Write a test for this run*, *Why did it fail?*, *Open the report*, or *Rerun the cases that failed*. Click one or press Alt-1 to Alt-3; suggestions that start work are put in the prompt first.
+**Next steps.** When a request finishes, up to three suggestions appear above the prompt: from what it produced (*Open the report*, *Write a test for this run*, *Why did it fail?*, *Rerun the cases that failed*) and from the agent's answer, which suggests the follow-ups you are likely to want (*Write a release report*, *Show the failed CD logs*). Click one or press Alt-1 to Alt-3; suggestions that start work are put in the prompt first, to adjust and send.
 
 | Key | Action |
 |-----|--------|

package/docs/user-guide.html

@@ -286,7 +286,7 @@ forgeqe docs --copy ./docs   # copy the guide, the architecture page, and the re
 <div class="area"><a class="area-name" href="#area-pull-requests"><span class="mono">Pull requests</span></a><ul class="tiles"><li><a class="tile" href="#cap-pr-review"><b>Review pull requests</b><span>Review PR 42 and focus on error handling</span></a></li><li><a class="tile" href="#cap-pr-impact"><b>Pull request impact and test selection</b><span>What does PR 42 impact?</span></a></li></ul></div>
 <div class="area"><a class="area-name" href="#area-api"><span class="mono">API</span></a><ul class="tiles"><li><a class="tile" href="#cap-api-testing"><b>API and contract testing</b><span>Turn the last run's API calls into tests</span></a></li></ul></div>
 <div class="area"><a class="area-name" href="#area-quality"><span class="mono">Quality</span></a><ul class="tiles"><li><a class="tile" href="#cap-accessibility-performance"><b>Accessibility and page performance</b><span>Is the checkout page accessible?</span></a></li><li><a class="tile" href="#cap-visual-comparison"><b>Visual comparison with baselines</b><span>What changed visually since the last run?</span></a></li></ul></div>
-<div class="area"><a class="area-name" href="#area-release"><span class="mono">Release</span></a><ul class="tiles"><li><a class="tile" href="#cap-release-readiness"><b>Requirement coverage and release readiness</b><span>Are we ready to release?</span></a></li></ul></div>
+<div class="area"><a class="area-name" href="#area-release"><span class="mono">Release</span></a><ul class="tiles"><li><a class="tile" href="#cap-release-readiness"><b>Requirement coverage and release readiness</b><span>Are we ready to release?</span></a></li><li><a class="tile" href="#cap-reports"><b>Reports you can share</b><span>Create a release report for all projects</span></a></li></ul></div>
 <div class="area"><a class="area-name" href="#area-operations"><span class="mono">Operations</span></a><ul class="tiles"><li><a class="tile" href="#cap-monitoring"><b>Scheduled monitoring with alerts</b><span>Check the login flow every 15 minutes and alert Teams</span></a></li><li><a class="tile" href="#cap-notifications"><b>Teams and email notifications</b><span>Post the results to the QA channel</span></a></li><li><a class="tile" href="#cap-pipelines"><b>CI pipelines</b><span>Add a GitHub Actions workflow that runs tests/cases.yaml on every PR, and make it green</span></a></li></ul></div>
 <div class="area"><a class="area-name" href="#area-systems"><span class="mono">Systems</span></a><ul class="tiles"><li><a class="tile" href="#cap-systems"><b>Work with Azure DevOps, Jira, and Confluence</b><span>Show the open bugs in this sprint</span></a></li></ul></div>
 <div class="area"><a class="area-name" href="#area-setup"><span class="mono">Setup</span></a><ul class="tiles"><li><a class="tile" href="#cap-setup-help"><b>Setup, skills, and documentation</b><span>What can you do?</span></a></li></ul></div>
@@ -550,6 +550,9 @@ Add a GitHub Actions workflow that runs tests/cases.yaml on every PR, and make i
 <li><strong>Right (60%)</strong>: a status line with who is working, for how long, and a <strong>■ Stop</strong> button (a running suite adds a progress bar, <code>██████░░░░░░ 5/8</code>); then <strong>Plan</strong> (the agent's to-do list, then each browser step or suite case with ✓, ✗ and why, ◐ running), <strong>Activity</strong> (every tool call, browser step, and the agents' reasoning), <strong>Files</strong> (each file read ◉, written ✚, or edited ✎, newest first), <strong>Context</strong> (how full the model's context window is, and the skills, product knowledge, domain, and MCP servers in it), and <strong>Tokens</strong> (total, input, output, cached; no prices, which are in reports and <code>forgeqe models</code>).</li>
 </ul>
 <p>On a narrower terminal the chat fills the screen, and Ctrl-O switches between the chat and the same panels.</p>
+<p><strong>Buttons.</strong> Under the prompt: <strong>✦ Refine</strong> (Ctrl-R; it fills in once there is text to refine) and <strong>⌕ Search</strong> (Ctrl-P, the command palette); while a request runs, <strong>■ Stop</strong> (Esc) takes Refine's place. In the status line above the panels: <strong>▦ Panels</strong> (Shift-Tab) and <strong>⤢ Wide chat</strong>, which gives the conversation the whole width for reading a long answer; <strong>◧ Split view</strong>, under the prompt, brings the panels back. On a narrow terminal, <strong>◧ Details</strong> (Ctrl-O) and <strong>⇄ Profile</strong> (Tab) are under the prompt. Each is clickable.</p>
+<p><strong>Progress.</strong> While a request runs, the status line above the panels shows a bar for how far the agent's plan has got (<code>███████░░░░░ 2/3</code>), or the suite's cases while a suite runs.</p>
+<p><strong>Replies.</strong> The agent's answers lead with the answer, then short headed sections; lists keep their bullets in a column, so a wrapped line stays under its text instead of running back to the margin.</p>
 <p><strong>Moving between panels.</strong> Shift-Tab (or a click) moves the keys from the prompt to Plan, Activity, then Files; the focused panel's border lights up and shows its keys. Esc, or typing, goes back to the prompt.</p>
 <div class="tbl"><table>
 <thead>
@@ -566,7 +569,7 @@ Add a GitHub Actions workflow that runs tests/cases.yaml on every PR, and make i
 <p><strong>Command palette.</strong> Ctrl-P opens a palette of every action, command, profile, and example request; type a few letters to find one (&quot;rr&quot; finds <em>Refine my request</em>, &quot;auth&quot; finds the sign-in commands), ↑/↓ to choose, Enter to run it. Example requests are put in the prompt for you to edit, never sent straight away.</p>
 <p><strong><code>/</code> and <code>@</code> in the prompt.</strong> Typing <code>/</code> lists the commands that match; typing <code>@</code> suggests references: recent runs (<em>last run</em>, then each by its folder), files in the repository by name or path (<code>@Login</code> finds <code>pages/Login.ts</code>), and, after <code>@</code> and a number such as <code>@1761</code>, the Azure DevOps work item with its title (read-only). Tab completes, ↑/↓ choose, Esc closes the list.</p>
 <p><strong>Refine.</strong> Ctrl-R (or <em>Refine my request</em> in the palette) turns what you typed into a precise request before it is sent. A read-only agent checks it against this setup: what ForgeQE can do here, the systems that are set up, the project's sprints and test cases, and recent runs. It shows the refined request with what it <strong>checked</strong> ✓, what it <strong>assumed</strong> ~, and what is <strong>missing</strong> ?. Enter uses it, <code>e</code> puts it in the prompt to edit, Esc keeps yours. Refine never runs anything, writes files, or changes a work item.</p>
-<p><strong>Next steps.</strong> When a request finishes, up to three suggestions appear above the prompt, such as <em>Write a test for this run</em>, <em>Why did it fail?</em>, <em>Open the report</em>, or <em>Rerun the cases that failed</em>. Click one or press Alt-1 to Alt-3; suggestions that start work are put in the prompt first.</p>
+<p><strong>Next steps.</strong> When a request finishes, up to three suggestions appear above the prompt: from what it produced (<em>Open the report</em>, <em>Write a test for this run</em>, <em>Why did it fail?</em>, <em>Rerun the cases that failed</em>) and from the agent's answer, which suggests the follow-ups you are likely to want (<em>Write a release report</em>, <em>Show the failed CD logs</em>). Click one or press Alt-1 to Alt-3; suggestions that start work are put in the prompt first, to adjust and send.</p>
 <div class="tbl"><table>
 <thead>
 <tr><th>Key</th><th>Action</th></tr>
@@ -1559,6 +1562,28 @@ forgeqe visual approve --page /home  # make this run's /home the baseline (asks
 </section>
 </div>
 </article>
+<article class="spec feat" id="cap-reports">
+<div class="hd"><span class="mono">Release</span><h3>Reports you can share</h3></div>
+<p class="ds">Writes up what ForgeQE's agent found, such as release status across projects, open defects, or pipeline health, as a Markdown file and an HTML page in the reports folder, with a summary in the chat and links to open them.</p>
+<div class="facts">
+<div class="fact"><span class="mono">Ask it</span><ul class="asks"><li>Create a release report for all projects</li><li>Write up the open defects as a report I can share</li></ul></div>
+<div class="fact"><span class="mono">What it needs</span><ul class="needs"><li><span class="need req">Required</span><span>A live model profile in llms.json (an API key or Microsoft Entra ID sign-in)</span></li></ul></div>
+<div class="fact"><span class="mono">Limits</span><ul class="plainlist"><li>The report holds what the agent read with its tools in this request; it does not invent figures, and it does not publish the page anywhere unless the reports folder is served (notify.report_url).</li></ul></div>
+<div class="fact"><span class="mono">Tools</span><p class="tools"><code>write_report</code></p></div>
+</div>
+<div class="detail"><span class="mono">In detail</span>
+<section class="doc" id="doc-how_to_use-reports-from-the-agent">
+<h4>Reports from the agent</h4>
+<div class="prose"><p>Ask for a report and ForgeQE's agent writes one you can share: &quot;create a release report for all projects&quot;, &quot;write up the open defects&quot;, &quot;summarise this sprint's test results as a report&quot;. It gathers the facts with its tools (Azure DevOps, Jira, the runs it made), then saves the report with <code>write_report</code> as a Markdown file and an HTML page in the reports folder (<code>export.out_dir</code>, <code>.forgeqe/reports/</code> by default):</p>
+<pre><code>.forgeqe/reports/release-status-all-projects-20261009-001512/
+  report.md     # to edit, paste into a wiki, or attach
+  report.html   # to open or share; light and dark, tables scroll on a phone
+</code></pre>
+<p>The chat gets a short summary and a card with both paths as links, and <strong>Open the report</strong> appears as a next step. The report holds only what the agent read in that request. To post it, ask for it (&quot;post the report to the release channel&quot;) once <code>notify.channels</code> is set up; the reports folder is linked when it is served (<code>notify.report_url</code>).</p>
+</div>
+</section>
+</div>
+</article>
 </section>
 <section class="chapter" id="area-operations" aria-labelledby="area-operations-h">
 <header class="chead"><span class="mono">11</span><h2 id="area-operations-h">Operations</h2><p>Monitoring and CI pipelines</p></header>

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@capagents/forgeqe",
-  "version": "1.0.5",
+  "version": "1.0.7",
   "description": "ForgeQE: a browser agent CLI that runs goals and test cases in Chrome and turns passing runs into tests in your repo",
   "license": "MIT",
   "type": "module",

package/prompts/brain.md

@@ -163,4 +163,7 @@ Rules:
 - Never ask for an API token, API key, or password for Azure DevOps, GitHub, Jira, Confluence, TestRail, Zephyr, qTest, Xray, or a model provider in the chat. When one of them is not set up, or rejects the credentials (401, 403, "needs <NAME> in the environment"), tell the user to type /auth <system> in the shell (or run forgeqe auth login <system> in a terminal): it signs in through the browser and saves the token, and you get it on the next message. /auth <ENV_NAME> saves any other secret the same way. Website logins for browser runs are different: those go through the logins tool.
 - Never push except with push_branch, never to the default branch, and never force.
 - Stay inside the working directory, except to read ForgeQE's run and memory folders.
-- Replies are shown as markdown: use short lists, bold, and links where they help, and a table only for a few short columns. Keep replies short and concrete; test cases, tests, and reviews you were asked for are given in full, and only your own commentary is kept short. End every turn with the structured answer: reply (what you did and found, or your question) and outcome.
+- Replies are shown as markdown in a narrow chat pane, so write for scanning: one or two plain sentences with the answer first, then a short `##` heading per group if there is more than one, bullets of one line each, and bold only for the one word or figure that matters in a line (a status, a count), never whole sentences. Use a table only for two or three short columns. Keep replies short and concrete; test cases, tests, and reviews you were asked for are given in full, and only your own commentary is kept short.
+- When the user asks for a report, a write-up, or a summary to share ("create a release report", "write up the open defects"), gather the facts with your tools, then save it with write_report and reply with three to five lines of summary; the shell shows the report's paths as links, so do not repeat them. Never only paste a long report into the reply.
+- Do not end a reply with "If you want, I can…" options: put up to three follow-ups the user is likely to want next in the structured answer's next (a short label, and the request as the user would type it). They are shown as buttons.
+- End every turn with the structured answer: reply (what you did and found, or your question), outcome, and next.

package/src/brain/index.ts

@@ -8,7 +8,7 @@ import type { AppConfig } from "../config.ts"
 import { externalMcp, externalNotes, type ExternalPlan } from "../mcp/external.ts"
 import { MCP_NAME, TOOL_TIMEOUT_MS, localToolServer } from "../mcp/server.ts"
 import { TOOLS, environmentNotes, type ToolContext } from "../mcp/tools.ts"
-import { brainPrompt, BRAIN_SCHEMA, type BrainReport } from "./prompt.ts"
+import { brainPrompt, BRAIN_SCHEMA, cleanFollowUps, type BrainReport, type FollowUp } from "./prompt.ts"
 import { redactSecrets } from "../sessionLogins.ts"
 import { logEvent } from "../sessionLog.ts"
 import { loadSkills } from "../skills/registry.ts"
@@ -16,7 +16,7 @@ import { domainNotes } from "../domain.ts"
 
 export const BRAIN_AGENT = "forgeqe"
 
-export type BrainAnswer = { reply: string; outcome: "done" | "failed" | "needs_input" | "stopped"; error?: string }
+export type BrainAnswer = { reply: string; outcome: "done" | "failed" | "needs_input" | "stopped"; error?: string; next?: FollowUp[] }
 
 export type Brain = {
   model: string
@@ -173,7 +173,8 @@ export async function openBrain(opts: BrainOptions): Promise<Brain> {
       if (stopped || out.error === "stopped" || opts.signal?.aborted) return { reply: "", outcome: "stopped" }
       const reply = out.report?.reply?.trim() || out.text.trim()
       if (out.error && !out.report) return { reply, outcome: "failed", error: out.error }
-      return { reply, outcome: out.report?.outcome || (out.error ? "failed" : "done"), error: out.error }
+      const next = cleanFollowUps(out.report?.next)
+      return { reply, outcome: out.report?.outcome || (out.error ? "failed" : "done"), error: out.error, ...(next.length ? { next } : {}) }
     },
     abort: () => {
       stopped = true

package/src/brain/prompt.ts

@@ -6,14 +6,46 @@ export const brainPrompt = () => prompt("brain")
 export const BRAIN_SCHEMA = {
   type: "object",
   properties: {
-    reply: { type: "string", description: "What you did and found, or the question you need answered. Plain text, short." },
+    reply: { type: "string", description: "What you did and found, or the question you need answered: Markdown, answer first, short lines." },
     outcome: {
       type: "string",
       enum: ["done", "failed", "needs_input"],
       description: "done: the request was carried out and everything checked passed. failed: a run, test, or pipeline failed or could not be done. needs_input: you asked the user something.",
     },
+    next: {
+      type: "array",
+      maxItems: 3,
+      description: "Up to three follow-ups the user is likely to want next, shown as buttons; empty when there is no clear next step.",
+      items: {
+        type: "object",
+        properties: {
+          label: { type: "string", description: "Two to five words, e.g. Write a release report" },
+          request: { type: "string", description: "The request as the user would type it, complete on its own" },
+        },
+        required: ["label", "request"],
+      },
+    },
   },
   required: ["reply", "outcome"],
 } as const
 
-export type BrainReport = { reply?: string; outcome?: "done" | "failed" | "needs_input" }
+/** A follow-up the agent suggests: a button label, and the request it puts in the prompt. */
+export type FollowUp = { label: string; request: string }
+
+export type BrainReport = { reply?: string; outcome?: "done" | "failed" | "needs_input"; next?: FollowUp[] }
+
+/** The follow-ups worth showing: well-formed, short enough for a button, at most three, no repeats. */
+export function cleanFollowUps(next: unknown): FollowUp[] {
+  if (!Array.isArray(next)) return []
+  const seen = new Set<string>()
+  const out: FollowUp[] = []
+  for (const n of next) {
+    const label = typeof n?.label === "string" ? n.label.replace(/\s+/g, " ").trim() : ""
+    const request = typeof n?.request === "string" ? n.request.trim() : ""
+    if (!label || !request || label.length > 40 || seen.has(label.toLowerCase())) continue
+    seen.add(label.toLowerCase())
+    out.push({ label, request })
+    if (out.length === 3) break
+  }
+  return out
+}

package/src/capabilities/registry.ts

@@ -290,6 +290,18 @@ export const CAPABILITIES: Capability[] = [
     limits: ["Automatic notifications come from runs nobody is watching (forgeqe suite, forgeqe run, pipelines, the MCP server), not from the interactive shell, where the agent posts only when asked."],
     docs: [{ doc: "HOW_TO_USE.md", heading: "17. Notifications" }],
   },
+  {
+    id: "reports",
+    area: "Release",
+    title: "Reports you can share",
+    summary:
+      "Writes up what ForgeQE's agent found, such as release status across projects, open defects, or pipeline health, as a Markdown file and an HTML page in the reports folder, with a summary in the chat and links to open them.",
+    tools: ["write_report"],
+    examples: ["Create a release report for all projects", "Write up the open defects as a report I can share"],
+    requires: [{ kind: "live_model" }],
+    limits: ["The report holds what the agent read with its tools in this request; it does not invent figures, and it does not publish the page anywhere unless the reports folder is served (notify.report_url)."],
+    docs: [{ doc: "HOW_TO_USE.md", heading: "Reports from the agent" }],
+  },
   {
     id: "pipelines",
     area: "Operations",

package/src/mcp/reporttools.ts

@@ -0,0 +1,101 @@
+import { mkdirSync, writeFileSync } from "node:fs"
+import path from "node:path"
+
+import type { McpTool } from "./tools.ts"
+
+/**
+ * Reports the agent writes from what it found (release status, open defects, pipeline health): a Markdown file to
+ * edit or paste, and the same report as a page to open or share. The chat gets the summary and the links.
+ */
+
+/** "Release status: all projects" → "release-status-all-projects". */
+export function reportSlug(title: string): string {
+  return (
+    title
+      .toLowerCase()
+      .replace(/[^a-z0-9]+/g, "-")
+      .slice(0, 40)
+      .replace(/^-+|-+$/g, "") || "report"
+  )
+}
+
+const escapeHtml = (s: string) => s.replace(/&/g, "&amp;").replace(/</g, "&lt;").replace(/>/g, "&gt;").replace(/"/g, "&quot;")
+
+/** The report as a page: the Markdown rendered, with raw HTML in it shown as text so a report can never run script. */
+export function reportHtml(title: string, markdown: string, written: Date): string {
+  const safe = markdown.replace(/<(?=[A-Za-z/!?])/g, "&lt;")
+  const body = Bun.markdown.html(safe)
+  return `<!doctype html>
+<html lang="en">
+<head>
+<meta charset="utf-8">
+<meta name="viewport" content="width=device-width, initial-scale=1">
+<title>${escapeHtml(title)}</title>
+<style>
+:root { --bg: #ffffff; --text: #1f2328; --muted: #59636e; --border: #d1d9e0; --accent: #0969da; --panel: #f6f8fa; }
+@media (prefers-color-scheme: dark) { :root { --bg: #0d1117; --text: #e6edf3; --muted: #8b949e; --border: #30363d; --accent: #4ea1ff; --panel: #161b22; } }
+* { box-sizing: border-box; }
+body { margin: 0; background: var(--bg); color: var(--text); font: 15px/1.6 -apple-system, "Segoe UI", Roboto, sans-serif; }
+main { max-width: 960px; margin: 0 auto; padding: 32px 16px 64px; }
+header { border-bottom: 1px solid var(--border); padding-bottom: 12px; margin-bottom: 24px; }
+header p { margin: 4px 0 0; color: var(--muted); font-size: 13px; }
+h1 { font-size: 28px; line-height: 1.25; margin: 0; }
+h2 { font-size: 20px; margin: 32px 0 8px; padding-bottom: 4px; border-bottom: 1px solid var(--border); }
+h3 { font-size: 16px; margin: 24px 0 8px; }
+a { color: var(--accent); }
+code { background: var(--panel); border-radius: 4px; padding: 1px 4px; font-size: 13px; }
+pre { background: var(--panel); border: 1px solid var(--border); border-radius: 6px; padding: 12px; overflow-x: auto; }
+pre code { background: none; padding: 0; }
+table { border-collapse: collapse; width: 100%; margin: 12px 0; display: block; overflow-x: auto; }
+th, td { border: 1px solid var(--border); padding: 6px 10px; text-align: left; vertical-align: top; }
+th { background: var(--panel); }
+blockquote { margin: 12px 0; padding: 4px 12px; border-left: 4px solid var(--border); color: var(--muted); }
+li { margin: 2px 0; }
+</style>
+</head>
+<body>
+<main>
+<header>
+<h1>${escapeHtml(title)}</h1>
+<p>Written by ForgeQE on ${escapeHtml(written.toLocaleString())}</p>
+</header>
+${body}
+</main>
+</body>
+</html>
+`
+}
+
+export const REPORT_TOOLS: McpTool[] = [
+  {
+    name: "write_report",
+    description:
+      'Save a report the user asks for ("create a release report", "write up the open defects", "give me a report I can share") as a Markdown file and an HTML page in the reports folder. Gather the facts with your other tools first, then pass the whole report as Markdown: a one-paragraph summary first, then ## sections, tables for per-item status (one row per project, case, or build), and links to builds, work items, and pull requests. Do not invent data you did not read. The user is shown the report paths as links, so reply with three to five lines of summary only: do not repeat the paths or paste the report.',
+    inputSchema: {
+      type: "object",
+      properties: {
+        title: { type: "string", description: "The report's title, e.g. Release status: all projects, 9 October 2026" },
+        markdown: { type: "string", description: "The report body in Markdown, without the title (it is added as the heading)" },
+      },
+      required: ["title", "markdown"],
+    },
+    async run(args, ctx) {
+      const title = String(args.title || "Report").trim()
+      const markdown = String(args.markdown || "").trim()
+      if (!markdown) throw new Error("The report is empty: pass its body as markdown.")
+      const written = new Date()
+      const stamp = written.toISOString().replace(/[-:]/g, "").replace("T", "-").slice(0, 15)
+      // Beside the run reports, so the shell links it and outDir publishing (notify.report_url) serves it too.
+      const dir = path.resolve(ctx.config.export.outDir, `${reportSlug(title)}-${stamp}`)
+      mkdirSync(dir, { recursive: true })
+      const md = path.join(dir, "report.md")
+      const html = path.join(dir, "report.html")
+      writeFileSync(md, `# ${title}\n\n${markdown}\n`)
+      writeFileSync(html, reportHtml(title, markdown, written))
+      const show = (p: string) => path.relative(process.cwd(), p) || p
+      ctx.emit?.({ type: "report", title, html, markdown: md })
+      ctx.emit?.({ type: "card", card: { status: "ok", title: `Report: ${title}`, body: [`Page      ${show(html)}`, `Markdown  ${show(md)}`] } })
+      return { html: show(html), markdown: show(md) }
+    },
+  },
+]

package/src/mcp/tools.ts

@@ -29,6 +29,7 @@ import { EXPLORE_TOOLS } from "./exploretools.ts"
 import { QUALITY_TOOLS } from "./qualitytools.ts"
 import { CAPABILITY_TOOLS } from "./captools.ts"
 import { NOTIFY_TOOLS } from "./notifytools.ts"
+import { REPORT_TOOLS } from "./reporttools.ts"
 import { capabilityIndex } from "../capabilities/render.ts"
 import { domainNotes } from "../domain.ts"
 import { readNetwork } from "../api/traffic.ts"
@@ -95,6 +96,8 @@ export type ToolEvent =
   | { type: "code-done"; result?: CodeResult; error?: string; ms: number }
   /** A tool's result to show as a panel (a test written, tests healed, a monitoring round). */
   | { type: "card"; card: Card }
+  /** write_report saved a report: its page and its Markdown. */
+  | { type: "report"; title: string; html: string; markdown: string }
 
 export type ToolContext = {
   config: AppConfig
@@ -722,6 +725,7 @@ export const TOOLS: McpTool[] = [
   ...QUALITY_TOOLS,
   ...CAPABILITY_TOOLS,
   ...NOTIFY_TOOLS,
+  ...REPORT_TOOLS,
 ]
 
 /** The setup the agent should know about: where it works, which systems are configured, and the defaults. */

package/src/tui/App.tsx

@@ -7,6 +7,7 @@ import { useEffect, useRef, useState } from "react"
 import { NO_MODEL_SET_UP } from "../llm.ts"
 import type { CodeEvent, ReplayMode } from "../types.ts"
 import { brainUnavailable, carryOver, openBrain, sessionSetup, type Brain, type BrainOptions } from "../brain/index.ts"
+import type { FollowUp } from "../brain/prompt.ts"
 import { INSTALL_HINT } from "../code/runtime.ts"
 import { briefText, detailFrom, shows, showsAgentTool, stepIntent, type Detail } from "../detail.ts"
 import { stepVerdictLines } from "../cases/types.ts"
@@ -23,7 +24,7 @@ import { counts } from "../suite/report.ts"
 import { buildSummary, shortPath } from "../summary.ts"
 import { meterSnapshot, requestLine } from "../usage.ts"
 import { cleanSelection, copyText } from "./clipboard.ts"
-import { ChoicePanel, Footer, Hints, Logo, Prompt, Spinner, type Hint } from "./components.tsx"
+import { ButtonBar, ChoicePanel, Footer, Hints, Logo, Prompt, Spinner, type Button, type Hint } from "./components.tsx"
 import { setupChecklist, type SetupItem } from "./setup.ts"
 import { EXAMPLES, TIPS, formatDuration, formatTokens, packageVersion, pick, shortCwd, theme } from "./theme.ts"
 import { emptyProduced, Line, Produced, producedCard, TestsSetting, TranscriptLine } from "./transcript.tsx"
@@ -48,6 +49,7 @@ import {
   nextFocus,
   planLines,
   PlanPanel,
+  planProgress,
   planRows,
   SessionPanel,
   StatusBar,
@@ -167,6 +169,8 @@ export function App({ options }: { options: TuiOptions }) {
   const requests = useRef(0)
   /** Under 120 columns there is room for one pane: ctrl+o switches between the chat and the details. */
   const [narrowView, setNarrowView] = useState<"chat" | "details">("chat")
+  // The split view's chat at the full width, for reading a long answer; the panels come back with Split view.
+  const [chatWide, setChatWide] = useState(false)
   const setupRef = useRef<{ key: string; items: SetupItem[] }>({ key: "", items: [] })
   // Which panel the keys go to (shift+tab moves on), the row selected in the plan and files, and the activity filter.
   const [nav, setNavState] = useState<{ focus: Focus; plan: number; files: number; filter: ActivityFilter }>({ focus: "chat", plan: 0, files: 0, filter: "all" })
@@ -481,6 +485,7 @@ export function App({ options }: { options: TuiOptions }) {
       active: configRef.current?.activeProfile || "",
       narrow: widthRef.current < SPLIT_MIN,
       view: narrowView,
+      chatWide,
     })
   const paletteMatches = (query: string) => searchPalette(query, paletteAll())
 
@@ -556,6 +561,8 @@ export function App({ options }: { options: TuiOptions }) {
     else if (run.id === "stop") openDialog("stop")
     else if (run.id === "details" || run.id === "chat") setNarrowView(run.id)
     else if (run.id === "refine") refineRef.current()
+    else if (run.id === "wide") setChatWide((w) => !w)
+    else if (run.id === "panels") setNav((n) => ({ ...n, focus: planLines(inspRef.current).length ? "plan" : "activity" }))
   }
 
   const cycleProfile = () => {
@@ -641,7 +648,7 @@ export function App({ options }: { options: TuiOptions }) {
       return
     }
     // The panels: shift+tab moves focus through those on screen; a focused panel takes the arrows, enter, and esc.
-    const panes = width >= SPLIT_MIN || narrowView === "details"
+    const panes = (width >= SPLIT_MIN && !chatWide) || (width < SPLIT_MIN && narrowView === "details")
     const insp = inspRef.current
     const shown = (f: Focus) => f === "activity" || f === "files" || (f === "plan" && planLines(insp).length > 0) || (f === "chat" && width >= SPLIT_MIN)
     if (panes && key.name === "tab" && key.shift && !entryRef.current) {
@@ -916,6 +923,9 @@ export function App({ options }: { options: TuiOptions }) {
       case "card":
         pushLine({ kind: "card", text: e.card.title, card: e.card })
         return
+      case "report":
+        producedRef.current.reports.push({ title: e.title, html: e.html })
+        return
     }
   }
 
@@ -971,8 +981,8 @@ export function App({ options }: { options: TuiOptions }) {
     return controller
   }
 
-  const finish = (finalStatus: string) => {
-    setNext(nextSteps(producedRef.current, { testsOn: testsRef.current.on }))
+  const finish = (finalStatus: string, followUps: FollowUp[] = []) => {
+    setNext(nextSteps(producedRef.current, { testsOn: testsRef.current.on, followUps }))
     progressDone(inspRef.current, finalStatus === "stopped")
     bump()
     setStatus(finalStatus)
@@ -1016,6 +1026,7 @@ export function App({ options }: { options: TuiOptions }) {
     const setup = sessionSetup(cfg, options.repo)
     testsRef.current = { on: setup.writeTests, why: "the session was started with --no-code" }
     let finalStatus = "fail"
+    let followUps: FollowUp[] = []
     let fresh = false
     try {
       if (!brainRef.current) {
@@ -1043,6 +1054,7 @@ export function App({ options }: { options: TuiOptions }) {
         finalStatus = "stopped"
       } else {
         if (answer.reply) pushLine({ kind: "reply", text: answer.reply, outcome: answer.outcome })
+        followUps = answer.next || []
         if (answer.error && !answer.reply) push("err", answer.error, "chat")
         finalStatus = answer.outcome === "done" ? "done" : answer.outcome === "needs_input" ? "idle" : "fail"
       }
@@ -1055,7 +1067,7 @@ export function App({ options }: { options: TuiOptions }) {
       // The split view's session and tokens panels show this; only the one-pane view needs it in the chat.
       const spent = widthRef.current < SPLIT_MIN ? requestLine(meterStart) : ""
       if (spent) push("sys", spent, "chat")
-      finish(finalStatus)
+      finish(finalStatus, followUps)
     }
   }
 
@@ -1285,35 +1297,38 @@ export function App({ options }: { options: TuiOptions }) {
 
   const closeHint: Hint = { key: "ctrl+c", label: "close", onPress: () => openDialog("quit") }
   const copiedHint: Hint[] = copied ? [{ key: `✓ ${copied}`, label: "" }] : []
-  // The split view's right pane shows tokens, time, and the last result, so the hints under the prompt leave them out.
+  // The split view's right pane shows tokens, time, and the last result, so the row under the prompt leaves them out.
   const wide = width >= SPLIT_MIN
-  const viewHint: Hint[] = wide ? [] : [{ key: "ctrl+o", label: "details", onPress: () => setNarrowView("details") }]
-  // Each hint's `drop` is when it gives way on a narrow row: the last run and profiles (also in the palette) go first.
-  const idleHints: Hint[] = [
-    ...copiedHint.map((h) => ({ ...h, drop: 0 })),
-    ...(lastStats && !wide ? [{ key: `last run ${lastStats}`, label: "", drop: 5 }] : []),
-    ...viewHint.map((h) => ({ ...h, drop: 3 })),
-    { key: "ctrl+r", label: "refine", onPress: () => refineRef.current(), drop: 1 },
-    ...(wide ? [{ key: "shift+tab", label: "panels", onPress: () => setNav((n) => ({ ...n, focus: planLines(inspRef.current).length ? "plan" : "activity" })), drop: 4 }] : [{ key: "tab", label: "profiles", drop: 4 }]),
-    { key: "ctrl+p", label: "commands" },
-    closeHint,
-  ]
-  const runningHints: Hint[] = [
-    // A note ("opened report.html") is the last to give way: it answers what was just pressed.
-    ...copiedHint.map((h) => ({ ...h, drop: 0 })),
-    // The split view's panels show tokens and time; the one-pane view shows them here while there is room.
-    ...(wide ? [] : [{ key: `${formatTokens(tokens)} tokens · ${elapsed}`, label: "", drop: 3 }, ...viewHint.map((h) => ({ ...h, drop: 2 }))]),
-    { key: "ctrl+p", label: "commands", drop: 1 },
-    closeHint,
+  const openPalette = () => {
+    paletteRef.current = { query: "", selected: 0 }
+    setPalette(paletteRef.current)
+  }
+  const focusPanels = () => setNav((n) => ({ ...n, focus: planLines(inspRef.current).length ? "plan" : "activity" }))
+  // What can be done from here, as buttons: Refine fills with its colour once there is a request to refine.
+  // `drop`: the order they give way on a narrow row; Refine, Search, and Stop always stay.
+  const refineButton: Button = { icon: "✦", label: "Refine", keys: "ctrl+r", color: theme.memory, onPress: () => refineRef.current(), active: Boolean(promptText.trim()) }
+  const searchButton: Button = { icon: "⌕", label: "Search", keys: "ctrl+p", color: theme.accent, onPress: openPalette }
+  // The view: in the split view, Panels and Wide chat sit in the right pane's status line, where there is room;
+  // Split view comes back here when the chat is wide, and Details on a narrow terminal.
+  const toggleWide = () => setChatWide((w) => !w)
+  const paneButtons: Button[] = [
+    { icon: "▦", label: "Panels", keys: "shift+tab", color: theme.thought, onPress: focusPanels, drop: 2 },
+    { icon: "⤢", label: "Wide chat", color: theme.muted, onPress: toggleWide, drop: 1 },
   ]
+  const viewButtons: Button[] = wide
+    ? chatWide
+      ? [{ icon: "◧", label: "Split view", color: theme.thought, onPress: toggleWide, drop: 2 }]
+      : []
+    : [{ icon: "◧", label: "Details", keys: "ctrl+o", color: theme.thought, onPress: () => setNarrowView("details"), drop: 1 }]
+  const idleButtons: Button[] = [refineButton, searchButton, ...viewButtons, ...(wide ? [] : [{ icon: "⇄", label: "Profile", keys: "tab", color: theme.muted, onPress: cycleProfile, drop: 4 }])]
+  const runningButtons: Button[] = [{ icon: "■", label: "Stop", keys: "esc", color: theme.warn, onPress: () => openDialog("stop") }, searchButton, ...viewButtons]
+  const idleHints: Hint[] = [...copiedHint, ...(lastStats && !wide && !copied ? [{ key: `last run ${lastStats}`, label: "" }] : []), closeHint]
+  // The one-pane view has no tokens panel, so the tokens and time show here.
+  const runningHints: Hint[] = [...copiedHint, ...(wide || copied ? [] : [{ key: `${formatTokens(tokens)} tokens · ${elapsed}`, label: "" }]), closeHint]
   const runningLeft = (
-    <>
+    <box style={{ flexDirection: "row", marginRight: 1 }}>
       <Spinner kind="bar" />
-      <text onMouseDown={() => openDialog("stop")}>
-        <span fg={theme.text}>{"  esc"}</span>
-        <span fg={theme.muted}>{" stop"}</span>
-      </text>
-    </>
+    </box>
   )
 
   const renderControls = (promptWidth: number, placeholder: string, rows: [number, number]) => {
@@ -1406,7 +1421,7 @@ export function App({ options }: { options: TuiOptions }) {
             setPromptText(t)
           }}
         />
-        <Hints width={promptWidth} items={running ? runningHints : idleHints} left={running ? runningLeft : undefined} reserve={running ? 20 : 0} />
+        <ButtonBar width={promptWidth} buttons={running ? runningButtons : idleButtons} hints={running ? runningHints : idleHints} left={running ? runningLeft : undefined} />
       </>
     )
   }
@@ -1509,7 +1524,7 @@ export function App({ options }: { options: TuiOptions }) {
     const sel = (f: "plan" | "files") => (nav.focus === f ? Math.min(nav[f], Math.max(0, (f === "plan" ? planLines(insp) : filesByUse(insp)).length - 1)) : undefined)
     return (
       <box style={{ flexDirection: "column", width: paneWidth, height: rowsAvailable, flexShrink: 0 }}>
-        <StatusBar running={running} who={insp.who} activity={activity} elapsed={elapsed} status={status} width={paneWidth} onStop={stop} progress={suiteProgress} />
+        <StatusBar running={running} who={insp.who} activity={activity} elapsed={elapsed} status={status} width={paneWidth} onStop={stop} progress={suiteProgress || planProgress(insp)} actions={paneButtons} />
         <box style={{ height: 1, flexShrink: 0 }} />
         {plan ? <PlanPanel s={insp} width={paneWidth} height={plan} selected={sel("plan")} onFocus={focusOn("plan")} /> : null}
         <ActivityPanel
@@ -1536,7 +1551,7 @@ export function App({ options }: { options: TuiOptions }) {
       {all.map((line) => (
         <TranscriptLine key={line.id} line={line} />
       ))}
-      {running && !split ? (
+      {running && (!split || chatWide) ? (
         <box style={{ flexDirection: "row", marginTop: 1, paddingLeft: 2, flexShrink: 0 }}>
           <Spinner kind="dots" />
           <text fg={theme.text}>{` ${activity || "Working…"}`}</text>
@@ -1548,18 +1563,22 @@ export function App({ options }: { options: TuiOptions }) {
 
   const split = width >= SPLIT_MIN
   if (split) {
-    const leftWidth = Math.floor(width * 0.4)
+    // Wide chat: the conversation across the screen for reading, the panels hidden until Split view.
+    const leftWidth = chatWide ? width : Math.floor(width * 0.4)
     return (
       <box style={{ flexDirection: "column", flexGrow: 1 }}>
         <box style={{ flexDirection: "row", height: height - 1, flexShrink: 0, paddingTop: 1 }}>
-          <box onMouseDown={() => navRef.current.focus !== "chat" && setNav((n) => ({ ...n, focus: "chat" }))} style={{ flexDirection: "column", width: leftWidth, height: height - 2, paddingLeft: 2, paddingRight: 1 }}>
-            <SessionPanel info={sessionInfo} width={leftWidth - 3} />
-            {transcript(chatLines, {})}
-            <box style={{ flexDirection: "column", flexShrink: 0 }}>{controls(leftWidth - 3)}</box>
-          </box>
-          <box style={{ flexDirection: "column", width: width - leftWidth, height: height - 2, paddingLeft: 1, paddingRight: 2 }}>
-            {details(width - leftWidth - 3, height - 2)}
+          <box onMouseDown={() => navRef.current.focus !== "chat" && setNav((n) => ({ ...n, focus: "chat" }))} style={{ flexDirection: "column", width: leftWidth, height: height - 2, paddingLeft: 2, paddingRight: chatWide ? 2 : 1 }}>
+            <SessionPanel info={sessionInfo} width={leftWidth - (chatWide ? 4 : 3)} />
+            {/* Wide chat shows what the agents do in the conversation, as the one-pane view does. */}
+            {transcript(chatWide ? lines : chatLines, {})}
+            <box style={{ flexDirection: "column", flexShrink: 0 }}>{controls(leftWidth - (chatWide ? 4 : 3))}</box>
           </box>
+          {chatWide ? null : (
+            <box style={{ flexDirection: "column", width: width - leftWidth, height: height - 2, paddingLeft: 1, paddingRight: 2 }}>
+              {details(width - leftWidth - 3, height - 2)}
+            </box>
+          )}
         </box>
         <Footer left={shortCwd()} right={footerRight} width={width} />
         {paletteOverlay}

package/src/tui/components.tsx

@@ -206,6 +206,55 @@ export function Hints({ width, items: all, left, reserve = 0 }: { width: number;
   )
 }
 
+/**
+ * A button under the prompt: an icon, a label, and its key, on a raised background; `active` fills it with its colour
+ * (Refine once there is text to refine). `drop`: when it gives way on a narrow row, highest first.
+ */
+export type Button = { icon: string; label: string; keys?: string; color: string; onPress: () => void; active?: boolean; drop?: number }
+
+const buttonCells = (b: Button) => b.icon.length + 1 + b.label.length + (b.keys ? b.keys.length + 1 : 0) + 3
+
+/** The buttons that fit in `room` columns; a narrow row first loses whole buttons by their `drop`, then the keys. */
+export function fitButtons(items: Button[], room: number): Button[] {
+  const cells = (list: Button[]) => list.reduce((n, b) => n + buttonCells(b), 0)
+  let kept = [...items]
+  for (const b of items.filter((x) => x.drop !== undefined).sort((x, y) => y.drop! - x.drop!)) {
+    if (cells(kept) <= room) return kept
+    kept = kept.filter((k) => k !== b)
+  }
+  return cells(kept) <= room ? kept : kept.map((b) => ({ ...b, keys: undefined }))
+}
+
+/** The row under the prompt: buttons on the left, the hints (close, a note) on the right. */
+export function ButtonBar({ width, buttons, hints, left }: { width: number; buttons: Button[]; hints: Hint[]; left?: ReactNode }) {
+  const hintCells = hints.reduce((n, h, i) => n + h.key.length + (h.label ? h.label.length + 1 : 0) + (i ? 2 : 0), 0)
+  const shown = fitButtons(buttons, Math.max(0, width - hintCells - (left ? 22 : 0) - 1))
+  return (
+    <box style={{ width, height: 1, flexShrink: 0, flexDirection: "row", justifyContent: "space-between" }}>
+      <box style={{ flexDirection: "row" }}>
+        {left}
+        {shown.map((b) => (
+          <box key={b.label} onMouseDown={b.onPress} style={{ flexDirection: "row", paddingLeft: 1, paddingRight: 1, marginRight: 1, backgroundColor: b.active ? b.color : theme.button }}>
+            <text>
+              <span fg={b.active ? theme.onButton : b.color}>{`${b.icon} `}</span>
+              <span fg={b.active ? theme.onButton : theme.text}>{b.label}</span>
+              {b.keys ? <span fg={b.active ? theme.onButton : theme.faint}>{` ${b.keys}`}</span> : null}
+            </text>
+          </box>
+        ))}
+      </box>
+      <box style={{ flexDirection: "row" }}>
+        {hints.map((item, i) => (
+          <text key={`${item.key}-${item.label}`} onMouseDown={item.onPress}>
+            <span fg={theme.text}>{`${i ? "  " : ""}${item.key}`}</span>
+            <span fg={item.onPress ? theme.muted : theme.faint}>{item.label ? ` ${item.label}` : ""}</span>
+          </text>
+        ))}
+      </box>
+    </box>
+  )
+}
+
 export type Choice = { id: string; label: string }
 
 export type ChoicePanelProps = {

package/src/tui/inspector.tsx

@@ -2,7 +2,7 @@ import type { ScrollBoxRenderable } from "@opentui/core"
 import type { ReactNode, RefObject } from "react"
 
 import type { Usage } from "../types.ts"
-import { Spinner } from "./components.tsx"
+import { fitButtons, Spinner, type Button } from "./components.tsx"
 import { filesByUse, type FileUse, type Inspector, type Progress } from "./session.ts"
 import { clip, tailCut } from "./text.ts"
 import { formatTokens, theme } from "./theme.ts"
@@ -102,6 +102,11 @@ export function SessionPanel({ info, width }: { info: SessionInfo; width: number
 }
 
 /** The right pane's first line: who is working on what and for how long, and Stop while something runs. */
+/** How far the agent's own plan has got, while it has one of two steps or more. */
+export function planProgress(s: Inspector): { done: number; total: number } | null {
+  return s.plan.length > 1 ? { done: s.plan.filter((p) => p.status === "completed").length, total: s.plan.length } : null
+}
+
 /** "███████░░░░░ 5/8": how far a suite has got, in `cells` cells. */
 export function progressBar(done: number, total: number, cells: number): [string, string] {
   const filled = total ? Math.round((Math.min(done, total) / total) * cells) : 0
@@ -118,6 +123,8 @@ export function StatusBar(props: {
   onStop: () => void
   /** A suite's cases: done so far, of how many. */
   progress?: { done: number; total: number } | null
+  /** Buttons for the view (Panels, Wide chat), before Stop; the first to give way when the line is short. */
+  actions?: Button[]
 }) {
   const done = props.status === "done"
   const failed = props.status === "fail"
@@ -134,7 +141,9 @@ export function StatusBar(props: {
   const bar = props.running && props.progress && props.progress.total > 1 ? props.progress : null
   const cells = 12
   const barText = bar ? ` ${bar.done}/${bar.total}  ` : ""
-  const room = Math.max(10, props.width - props.who.length - (props.running ? 30 : 6) - (bar ? cells + barText.length : 0))
+  const fixed = props.who.length + (props.running ? 30 : 6) + (bar ? cells + barText.length : 0)
+  const actions = fitButtons(props.actions || [], Math.max(0, props.width - fixed - 28))
+  const room = Math.max(10, props.width - fixed - actions.reduce((n, b) => n + b.icon.length + b.label.length + (b.keys ? b.keys.length + 1 : 0) + 4, 0))
   const [full, empty] = bar ? progressBar(bar.done, bar.total, cells) : ["", ""]
   return (
     <box style={{ flexDirection: "row", justifyContent: "space-between", height: 1, flexShrink: 0, paddingLeft: 1, paddingRight: 1 }}>
@@ -145,8 +154,18 @@ export function StatusBar(props: {
           <span fg={props.running ? theme.muted : color}>{`${props.running ? "  " : " "}${clip(what, room)}`}</span>
         </text>
       </box>
+      <box style={{ flexDirection: "row" }}>
+        {actions.map((b) => (
+          <box key={b.label} onMouseDown={b.onPress} style={{ flexDirection: "row", paddingLeft: 1, paddingRight: 1, marginLeft: 1, backgroundColor: theme.button }}>
+            <text>
+              <span fg={b.color}>{`${b.icon} `}</span>
+              <span fg={theme.text}>{b.label}</span>
+              {b.keys ? <span fg={theme.faint}>{` ${b.keys}`}</span> : null}
+            </text>
+          </box>
+        ))}
       {props.running ? (
-        <box style={{ flexDirection: "row" }}>
+        <box style={{ flexDirection: "row", marginLeft: 2 }}>
           {bar ? (
             <text>
               <span fg={theme.ok}>{full}</span>
@@ -160,6 +179,7 @@ export function StatusBar(props: {
           </box>
         </box>
       ) : null}
+      </box>
     </box>
   )
 }
@@ -242,6 +262,12 @@ const ACTIVITY_STYLE: Record<string, [string, string, string]> = {
   err: ["✗", theme.err, theme.text],
 }
 
+/** A tool line's call and its arguments in brackets, or null when it has none. */
+export function toolArgs(text: string): [string, string] | null {
+  const at = text.indexOf(" (")
+  return at > 0 && text.endsWith(")") ? [text.slice(0, at), text.slice(at + 1)] : null
+}
+
 /** One activity line, its mark in a column of its own so wrapped text stays under the text. */
 export function ActivityRow({ line }: { line: Line }) {
   const [mark, markColor, textColor] = ACTIVITY_STYLE[line.kind] || ["·", theme.faint, theme.muted]
@@ -250,9 +276,17 @@ export function ActivityRow({ line }: { line: Line }) {
       <text fg={markColor} style={{ width: 2, flexShrink: 0 }}>
         {mark}
       </text>
-      <text fg={textColor} style={{ flexGrow: 1, flexShrink: 1 }}>
-        {line.text || " "}
-      </text>
+      {line.kind === "tool" && toolArgs(line.text) ? (
+        // "Azure DevOps · pipelines build (action list, project QI-Claims)": the call readable, its arguments quiet.
+        <text style={{ flexGrow: 1, flexShrink: 1 }}>
+          <span fg={theme.text}>{toolArgs(line.text)![0]}</span>
+          <span fg={theme.faint}>{` ${toolArgs(line.text)![1]}`}</span>
+        </text>
+      ) : (
+        <text fg={textColor} style={{ flexGrow: 1, flexShrink: 1 }}>
+          {line.text || " "}
+        </text>
+      )}
     </box>
   )
 }

package/src/tui/next.ts

@@ -1,6 +1,7 @@
 import { existsSync } from "node:fs"
 import path from "node:path"
 
+import type { FollowUp } from "../brain/prompt.ts"
 import type { Produced } from "./transcript.tsx"
 
 /**
@@ -10,8 +11,14 @@ import type { Produced } from "./transcript.tsx"
 
 export type NextStep = { label: string; run: { kind: "insert"; text: string } | { kind: "open"; path: string } }
 
-export function nextSteps(p: Produced, opts: { testsOn: boolean }): NextStep[] {
+/**
+ * `followUps`: what ForgeQE's agent suggests from its answer ("Write a release report"), after what the request
+ * produced; a follow-up that repeats a step already offered is left out.
+ */
+export function nextSteps(p: Produced, opts: { testsOn: boolean; followUps?: FollowUp[] }): NextStep[] {
   const steps: NextStep[] = []
+  const report = p.reports.at(-1)
+  if (report && existsSync(report.html)) steps.push({ label: "Open the report", run: { kind: "open", path: report.html } })
   const run = p.runs.at(-1)
   const suite = p.suites.at(-1)
   if (suite) {
@@ -28,5 +35,9 @@ export function nextSteps(p: Produced, opts: { testsOn: boolean }): NextStep[] {
   }
   const test = p.tests.at(-1)
   if (test?.result?.testFile && test.result.status === "passed") steps.unshift({ label: "Open the new test", run: { kind: "open", path: test.result.testFile } })
+  for (const f of opts.followUps || []) {
+    if (steps.some((s) => s.label.toLowerCase() === f.label.toLowerCase())) continue
+    steps.push({ label: f.label, run: { kind: "insert", text: f.request } })
+  }
   return steps.slice(0, 3)
 }

package/src/tui/palette.tsx

@@ -15,17 +15,21 @@ export type PaletteRun =
   /** Put in the prompt to edit and send: an example request. */
   | { kind: "insert"; text: string }
   /** Something the shell does itself. */
-  | { kind: "action"; id: "refine" | "stop" | "details" | "chat" }
+  | { kind: "action"; id: "refine" | "stop" | "details" | "chat" | "wide" | "panels" }
 
 export type PaletteItem = { id: string; group: PaletteGroup; title: string; detail?: string; keys?: string; run: PaletteRun }
 
 export const GROUP_COLOR: Record<PaletteGroup, string> = { Action: theme.warn, Command: theme.accent, Profile: theme.memory, Ask: theme.ok }
 
-export function paletteItems(opts: { running: boolean; headed: boolean; detailsFull: boolean; profiles: string[]; active: string; narrow: boolean; view: "chat" | "details" }): PaletteItem[] {
+export function paletteItems(opts: { running: boolean; headed: boolean; detailsFull: boolean; profiles: string[]; active: string; narrow: boolean; view: "chat" | "details"; chatWide?: boolean }): PaletteItem[] {
   const items: PaletteItem[] = []
   const command = (id: string, title: string, text: string, detail?: string, keys?: string) => items.push({ id, group: "Command", title, detail, keys, run: { kind: "command", text } })
   if (opts.running) items.push({ id: "stop", group: "Action", title: "Stop the running request", keys: "esc", run: { kind: "action", id: "stop" } })
   items.push({ id: "refine", group: "Action", title: "Refine my request", detail: "check it against what ForgeQE knows before sending", keys: "ctrl+r", run: { kind: "action", id: "refine" } })
+  if (!opts.narrow) {
+    items.push({ id: "wide", group: "Action", title: opts.chatWide ? "Split view: chat and panels" : "Wide chat: the conversation across the screen", run: { kind: "action", id: "wide" } })
+    if (!opts.chatWide) items.push({ id: "panels", group: "Action", title: "Move to the panels: plan, activity, files", keys: "shift+tab", run: { kind: "action", id: "panels" } })
+  }
   if (opts.narrow) items.push({ id: "view", group: "Action", title: opts.view === "chat" ? "Show the details view" : "Back to the chat", keys: "ctrl+o", run: { kind: "action", id: opts.view === "chat" ? "details" : "chat" } })
   command("features", "What ForgeQE can do here", "/features", "and what each feature still needs")
   command("auth", "Sign in to a system", "/auth", "Azure DevOps, GitHub, Jira, Confluence, TestRail, Zephyr, qTest, Xray")

package/src/tui/transcript.tsx

@@ -31,9 +31,11 @@ export type Produced = {
   runs: { runDir: string; status: RunStatus; files: string[] }[]
   suites: { html: string; junit: string }[]
   tests: { runDir: string; result?: CodeResult; error?: string }[]
+  /** Reports write_report saved: the page to open. */
+  reports: { title: string; html: string }[]
 }
 
-export const emptyProduced = (): Produced => ({ runs: [], suites: [], tests: [] })
+export const emptyProduced = (): Produced => ({ runs: [], suites: [], tests: [], reports: [] })
 
 /** Whether passed runs become tests this session, and why not when they don't. */
 export type TestsSetting = { on: boolean; why: string }
@@ -325,6 +327,80 @@ function replyStyle(): SyntaxStyle {
   return style
 }
 
+/** A reply cut into what reads better laid out than wrapped: headings, list items with their depth, and the rest. */
+export type ReplyBlock =
+  | { kind: "heading"; text: string }
+  | { kind: "item"; depth: number; mark: string; text: string }
+  | { kind: "markdown"; text: string }
+
+const ITEM = /^(\s*)([-*+]|\d+[.)])\s+(.*)$/
+
+export function replyBlocks(md: string): ReplyBlock[] {
+  const blocks: ReplyBlock[] = []
+  let rest: string[] = []
+  let fence = false
+  const flush = () => {
+    const text = rest.join("\n").trim()
+    if (text) blocks.push({ kind: "markdown", text })
+    rest = []
+  }
+  for (const line of md.replace(/\r/g, "").split("\n")) {
+    if (/^\s*(```|~~~)/.test(line)) fence = !fence
+    const last = blocks.at(-1)
+    if (fence || /^\s*(```|~~~)/.test(line) || /^\s*\|/.test(line)) {
+      rest.push(line)
+      continue
+    }
+    const heading = /^#{1,6}\s+(.*)$/.exec(line)
+    const item = ITEM.exec(line)
+    if (heading) {
+      flush()
+      blocks.push({ kind: "heading", text: heading[1]!.replace(/\*\*/g, "").trim() })
+    } else if (item) {
+      flush()
+      const depth = Math.min(3, Math.floor(item[1]!.replace(/\t/g, "  ").length / 2))
+      const ordered = /\d/.test(item[2]!)
+      blocks.push({ kind: "item", depth, mark: ordered ? item[2]!.replace(")", ".") : depth ? "◦" : "•", text: item[3]! })
+    } else if (line.trim() && /^\s+\S/.test(line) && last?.kind === "item" && !rest.length) {
+      // A list item's continuation line: it belongs to the item, under its text.
+      last.text += ` ${line.trim()}`
+    } else if (!line.trim()) {
+      rest.push(line)
+    } else rest.push(line)
+  }
+  flush()
+  return blocks
+}
+
+/** The agent's reply: headings with room above, list items with their marks in a column so wrapped lines hang under the text. */
+function ReplyBody({ text }: { text: string }) {
+  const blocks = replyBlocks(text)
+  return (
+    <box style={{ flexDirection: "column", flexShrink: 0 }}>
+      {blocks.map((b, i) =>
+        b.kind === "heading" ? (
+          <text key={i} fg={theme.accent} style={{ marginTop: i ? 1 : 0, flexShrink: 0 }}>
+            <strong>{b.text}</strong>
+          </text>
+        ) : b.kind === "item" ? (
+          <box key={i} style={{ flexDirection: "row", flexShrink: 0, paddingLeft: b.depth * 2, marginTop: blocks[i - 1]?.kind === "markdown" ? 1 : 0 }}>
+            <text fg={theme.accent} style={{ width: b.mark.length + 1, flexShrink: 0 }}>
+              {b.mark}
+            </text>
+            <box style={{ flexGrow: 1, flexShrink: 1 }}>
+              <markdown content={b.text} syntaxStyle={replyStyle()} fg={theme.text} conceal />
+            </box>
+          </box>
+        ) : (
+          <box key={i} style={{ marginTop: i ? 1 : 0, flexShrink: 0 }}>
+            <markdown content={b.text} syntaxStyle={replyStyle()} fg={theme.text} conceal tableOptions={{ widthMode: "content", wrapMode: "word" }} />
+          </box>
+        ),
+      )}
+    </box>
+  )
+}
+
 export function TranscriptLine({ line }: { line: Line }) {
   if (line.kind === "card" && line.card) return <ResultCard card={line.card} />
   if (line.kind === "summary" && line.summary) return <SummaryCard summary={line.summary} />
@@ -338,7 +414,7 @@ export function TranscriptLine({ line }: { line: Line }) {
         customBorderChars={BAR}
         style={{ backgroundColor: theme.panel, paddingLeft: 2, paddingRight: 2, paddingTop: 1, paddingBottom: 1, marginTop: 1, marginBottom: 1, flexShrink: 0 }}
       >
-        <markdown content={line.text} syntaxStyle={replyStyle()} fg={theme.text} conceal tableOptions={{ widthMode: "content", wrapMode: "word" }} />
+        <ReplyBody text={line.text} />
       </box>
     )
   }