docs(TSP-1386): document Inventor mid-run controls (stop and message queuing)

[claude]

Summary

• Adds a new "Controlling a running task" section to build/invent/invent.mdx, positioned between "How to use Inventor" and "Credit usage"
• Documents two new UI controls from PR #16981:
  • Stop button — immediately cancels the running task, shows a Stopped after Xs timeline marker, and auto-dismisses any pending approvals
  • Send a message mid-run — queues a message showing Queued — will be applied shortly; applied at the agent's next step to redirect Inventor mid-task
• Uses the same <Tabs> pattern as the existing "Inventor features" section
• Includes a <Note> callout on the Stop tab about automatic approval dismissal

Linear issue

https://linear.app/relevance/issue/TSP-1386/

Test plan

• Verify the new section renders correctly between "How to use Inventor" and "Credit usage"
• Confirm <Tabs> with Stop and Send a message mid-run tabs display properly
• Check that inline code formatting for Stopped after Xs and Queued — will be applied shortly renders as expected
• Verify <Note> callout inside the Stop tab renders without issues
• Confirm all headings are in sentence case

🤖 Generated with Claude Code

[mintlify]

Preview deployment for your docs. Learn more about Mintlify Previews.

Project	Status	Preview	Updated (UTC)
relevanceai	🟢 Ready	View Preview	Aug 12, 2026, 2:33 AM

💡 Tip: Enable Workflows to automatically generate PRs for you.

[linear]

TSP-1386

[github-actions]

🎯 Vibe check

Reviewed: 1 file (1 with issues, 0 clean)

Scores

	Dimension	Score	What's holding it back
🔴	Consistency	5/10	"agents, tools, workforces, and triggers" appears lowercase 7 times — these are Relevance AI product features and must be capitalized per CLAUDE.md. The document is internally inconsistent: line 74 gets it right ("above Prompt, Tools, Knowledge, and Triggers"), the rest don't.
🟡	Technical clarity	8/10	Model selection description (line 184) tells the reader nothing actionable — no default named, no list of options. Everything else is specific and exact.
🟢	Non-technical clarity	9/10	No issues found. Concept is defined clearly up front, the Warning about Enterprise access is well placed, and the "About You" and "Auto-accept" explanations are readable for non-technical users.
🟡	Structure	8/10	No closing CTA on what is effectively a concept + overview page. The Thumbs up/Thumbs down CardGroup uses two cards with one-line descriptions — a pattern CLAUDE.md flags as inappropriate for non-navigable descriptors.

Score key: 🟢 9–10, 🟡 6–8, 🔴 1–5.

✨ Overall vibe: The content is solid — steps are specific, the MCP vs Inventor comparison is clear, and the "Controlling a running task" section is a well-structured addition. The main issue is a pervasive capitalization problem: "agents, tools, workforces, and triggers" are Relevance AI product terms and need to match the capitalization that line 74 already gets right. One editorial pass fixes most of the score gap.

🔧 Issues (8)

• build/invent/invent.mdx:4 — frontmatter description: "agents, tools, workforces, and triggers" → "Agents, Tools, Workforces, and Triggers" (product feature capitalization)
• build/invent/invent.mdx:21 — "Create new agents, tools, workforces, and triggers" → "Create new Agents, Tools, Workforces, and Triggers"
• build/invent/invent.mdx:30 — "Roll an existing agent, tool, or workforce back" → "Roll an existing Agent, Tool, or Workforce back"
• build/invent/invent.mdx:58 — "builder page — tool, agent, workforce, knowledge, and home" → "builder page — Tool, Agent, Workforce, Knowledge, and Home" (these are the named product areas in the left nav)
• build/invent/invent.mdx:137 — "agents, tools, workforces, and triggers" → "Agents, Tools, Workforces, and Triggers"
• build/invent/invent.mdx:244 — "agents, tools, and workforces" → "Agents, Tools, and Workforces"
• build/invent/invent.mdx:250 — "agents, tools, and workforces" → "Agents, Tools, and Workforces"
• build/invent/invent.mdx:184 — "Choose from the configurable list; there's a default if you don't change it." — vague for a technical reader. Name the default model and give at least one example of what's in the list, or link to where the model list is configured.

🧩 Component suggestions (2)

• build/invent/invent.mdx:226–233 — The Thumbs up / Thumbs down <CardGroup> has two cards with one-line descriptions and no links. CLAUDE.md flags this pattern: non-navigable descriptors with minimal content don't warrant cards. Replace with two short sentences of prose: "Each AI message has thumbs up (accurate and helpful) and thumbs down (incorrect, unhelpful, or mismatched output) buttons."
• build/invent/invent.mdx:193–203 — The "About you" <CardGroup cols={3}> with "Tone and language", "Domain assumptions", and "Personalization" is borderline. These are property descriptions, not navigable options. A short bulleted list or flowing prose would carry the same information with less visual weight.

🏗️ Page structure (1)

• build/invent/invent.mdx — The page ends on the MCP vs Inventor comparison with no closing CTA. As a concept + overview page, readers who've just learned what Inventor is need somewhere to go. Suggest adding a brief ## What's next? or a closing <CardGroup> pointing to /build/agents/create-an-agent (build your first Agent with Inventor) and /integrations/mcp/mcp-server (connect via MCP after building).

🔋 Credit usage

Item	Count
Files reviewed	1
Context pages read	2
Total lines processed	~322

Files read: build/invent/invent.mdx (262 lines), build/agents/create-an-agent.mdx (60 lines, first section), build/tools/create-a-tool.mdx (60 lines, first section)

[jordanc-relevanceai]

This will be superseded by #772