fix(managed_agents): render outputs on docs site, retitle multiagent#600
Merged
Conversation
- Retitle CMA_coordinate_specialist_team to "Multiagent: coordinate a
specialist team" to match the Feature: pattern of sibling tutorials
- Lowercase the Outcomes title body ("Outcomes: agents that verify their
own work") to keep both Feature: titles on the same convention
- Replace display(Markdown(...)) with print() in the review loop and
brief/proposal extraction cells; the docs site does not render
IPython rich-display output, so these cells appeared empty
- Show section structure / sources rather than full prose in the final
cells to keep rendered output compact
Notebook ChangesThis PR modifies the following notebooks: 📓
|
![github-actions[bot]](https://avatars.githubusercontent.com/in/15368?s=60&v=4){kind=small}
There was a problem hiding this comment.
PR Review
Recommendation: APPROVE
Summary
Follow-up fixes that swap display(Markdown(...)) for print() so the review-loop and brief/proposal extraction cells render correctly on the docs site, plus title-casing normalization across both notebooks and registry.yaml.
Actionable Feedback (0 items)
No blocking or optional items. Diff is tight and self-consistent.
Detailed Review
Code Quality
- Section-extraction logic is correct.
line.startswith(\"#\")matches all H1/H2/H3 headers in the proposal (including the### St. Clair Health/### BlueRidge Health Planrows shown in the recorded output). For the brief,line.startswith((\"#\", \"[\"))correctly captures both section headings and the[N] \"...\"source entries — the diff shows all 6 sources came through cleanly. - No leftover IPython imports. Both
from IPython.display import Markdown, displaylines were removed; nothing else in either notebook still referencesIPythonordisplay(Markdown. - Outputs match the new code. The new
stdoutblocks line up with what the rewritten loops would produce.
registry.yaml
'Multiagent: coordinate a specialist team'matches the H1 at the top ofCMA_coordinate_specialist_team.ipynb.'Outcomes: agents that verify their own work'matches the H1 at the top ofCMA_verify_with_outcome_grader.ipynb.- Lowercase-after-colon convention now applied consistently across both
Feature:siblings.
Narrative consistency
- §5 prose in both notebooks was updated symmetrically: "look at the sections / cited sources... Print
proposal(orcontent) itself if you want to read the full document." Nice parallel construction, and the escape hatch preserves the original capability for readers who want the full prose. - §4 intros ("render each phase as it happens", "draws a labeled divider") still describe the rewritten helpers accurately — the banner still draws a labeled divider, just via
printnow.
Information trade-off
Well-judged. For §5, the full prose was already in scope in proposal / content, so pointing readers at those variables preserves access while keeping the rendered docs page scannable. Section headers + cited sources are the right summary view for a research brief — they answer "did the rubric coverage land?" at a glance. The §4 loop loses bold emphasis and the ⟳ / ✓ glyphs lose visual weight, but unicode dividers carry enough structure in monospace that the loop stays readable.
Positive Notes
- Tight, focused diff: pure rendering/title fix, no scope creep.
- The PR description clearly identifies the upstream pipeline gap (notebook → MDX converter only handling
streamoutputs) and notes the same break exists inclaude_agent_sdk/01_The_chief_of_staff_agent.ipynb, which is useful context for follow-up work. - Symmetry across the two notebooks (matching extraction patterns and matching prose hint) makes the change easy to maintain.
PedramNavid
approved these changes
May 6, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Follow-up fixes to the multiagent and outcomes cookbooks (#599):
Rendering on the docs site. The docs site's notebook → MDX converter only renders
streamoutputs, notdisplay_datawithtext/markdown. The review loop, final brief, and sales proposal cells all useddisplay(Markdown(...))and so showed empty output boxes on platform.claude.com. (This is a pre-existing pipeline gap —claude_agent_sdk/01_The_chief_of_staff_agent.ipynbhas had the same break since Sept 2025.) Fix:print()with plain-text dividers; output converted to a single stream blockTitle normalization.
# Building a Sales Proposal with a Heterogeneous Agent Team→# Multiagent: coordinate a specialist team, and# Outcomes: Agents that verify their own work→# Outcomes: agents that verify their own work, so bothFeature:siblings use the same lowercase-after-colon convention. Same change inregistry.yaml. URL slugs are path-derived so docs URLs are unchanged.Public port of anthropics/claude-cookbooks-private#67.