Anthropic Just Dropped a Masterclass on Building Agent Harnesses (for Large Codebases)
Walking through Anthropic's blog post on using Claude Code in large code bases, this video demos each strategy in a real repo: lean and layered CLAUDE.md files, self-improving stop/start hooks, path-scoped skills, an LSP-backed MCP server for symbol-level search, and subagents that split exploration from editing.
Cole MedinWatchTranscript found
Quick learning frame
Read this before watching.
Coding-agent workflow is the loop of inspect, route, plan, edit, verify, summarize, and decide what should be automated next.
New playlist item from Cole Medin; queued for transcript-backed review, topic mapping, and a practical learning artifact.
Skill you build: The ability to build an 'AI layer' of scoped context and tools around a coding agent so it can navigate and edit a large code base effectively, curating context up front instead of relying on model quality alone.
Watch for the shift from claim to mechanism. The learning value is the point where the transcript reveals a repeatable action, tool boundary, context move, review habit, or artifact.
Concept diagram
Where this video fits.
01Inspect context
02Route tool
03Plan work
04Edit safely
05Verify behavior
06Report next step
Deep lesson
Turn this video into working knowledge.
5,821 cleaned transcript words reviewed across 1,638 timed caption segments.
Thesis
Anthropic Just Dropped a Masterclass on Building Agent Harnesses (for Large Codebases) teaches a practical coding-agent workflow move: Walking through Anthropic's blog post on using Claude Code in large code bases, this video demos each strategy in a real repo: lean and layered CLAUDE.md files, self-improving stop/start hooks, path-scoped skills, an LSP-backed MCP server for symbol-level search, and subagents that split exploration from editing.
The goal is not to remember the video. The goal is to extract the operating principle, tie it to timestamped evidence, test how far the claim transfers, and make something reusable.
0:15
Harness over model
“already have a complex code base or two or three. You've got the apps and platforms that you're building, your second brain. You have these code bases that are tens or even hundreds of thousands of lines long...”
Claude Code uses agentic search (grep and folder structure, like an engineer) rather than RAG or code-base indexing, so there's no index to keep in sync but it works best with enough starting context; the core thesis is that the harness (the AI layer of context and tools) matters as much as the model. Sketch the three parts of your own code base as code, tests, and AI layer, and list which Claude Code features (rules, skills, hooks, MCP, subagents, LSP) you currently have in that third layer.
15:35
Self-improving hooks
“workflows and capabilities. And so, like this is an example of a skill right here for adding API routes in this code base. Really, a skill is some kind of set of steps, some kind of process, reusable...”
A stop hook runs a separate Claude session in headless mode when a turn ends to reflect on the changes and propose CLAUDE.md updates into a markdown review file, while a start hook loads dynamic context (like git status and recent commits, or Confluence docs), keeping global rules from going stale. Add a stop hook to one repo that spawns a headless Claude to review your diff against your CLAUDE.md and write suggested rule updates to a markdown file you can action later.
23:31
LSP beats grep
“authentication." I don't know, I'm just kind of throwing out something off the cuff here, but you have sub agents built into a lot of these coding agents now, like Claude Code and CodeX. And so, you don't...”
For six-digit line-count code bases grep is slow and token-inefficient, so a local MCP server exposing a language server protocol gives Claude the same where-is and find-references navigation an engineer has in their IDE, returning precise definitions and references by symbol instead of by string. Try a symbol-level search in a large repo by prompting Claude to find every reference to one specific function without using grep, forcing it to use LSP-style find-references tools.
01
Inspect context
Start with this video's job: Walking through Anthropic's blog post on using Claude Code in large code bases, this video demos each strategy in a real repo: lean and layered CLAUDE.md files, self-improving stop/start hooks, path-scoped skills, an LSP-backed MCP server for symbol-level search, and subagents that split exploration from editing. Treat "Inspect context" as the outcome you are trying to make visible, not a topic label. Anchor it to 0:15, where the video says: “already have a complex code base or two or three. You've got the apps and platforms that you're building, your second brain. You have these code bases that are tens or even hundreds of thousands of lines long...”
02
Route tool
Use "Route tool" to locate the part of the coding-agent workflow mechanism the video is demonstrating. Ask what changes in your real setup if this claim is true. Anchor it to 15:35, where the video says: “workflows and capabilities. And so, like this is an example of a skill right here for adding API routes in this code base. Really, a skill is some kind of set of steps, some kind of process, reusable...”
03
Plan work
Turn "Plan work" into the reusable artifact for this lesson: A coding-agent routing and execution matrix with context needed, tool choice, verification, and done signal. This is where watching becomes something you can inspect and reuse.
04
Edit safely
Use "Edit safely" as the application surface. Decide whether the idea touches a browser flow, a local file, a model choice, a source document, a UI, or a review step.
05
Verify behavior
Use "Verify behavior" to prove the lesson. The evidence should connect back to the video title, transcript anchors, and a concrete output, not a generic best-practice claim.
06
Report next step
Use "Report next step" to carry the idea forward: save the prompt, checklist, diagram, or operating rule that would make the next agent run better.
Example
Source-backed artifact packet
Convert the video into a scoped artifact request that includes the transcript claim, mechanism, acceptance criteria, and proof. The output should be a coding-agent routing and execution matrix with context needed, tool choice, verification, and done signal..
Example
Coding-agent workflow proof brief
Separate what the speaker claims, what the demo actually proves, and what still needs outside verification before you adopt the coding-agent workflow pattern.
Example
Teach-back module
Transform the lesson into a definition, a Inspect context -> Route tool -> Plan work -> Edit safely -> Verify behavior -> Report next step diagram, one misconception, one practice exercise, and a check-for-understanding question.
Do not learn it wrong
Treating the title as the lesson without checking what the transcript actually says.
choosing tools by hype
losing context across agents
letting parallel sessions become invisible
Letting the lesson drift into generic Codex vs Claude comparison.
Letting the lesson drift into feature lists without task routing.
Letting the lesson drift into claims that ignore limits or recovery.
Do not count this as learned until these are true.
01
State the transcript-backed claim in your own words: Walking through Anthropic's blog post on using Claude Code in large code bases, this video demos each strategy in a real repo: lean and layered CLAUDE.md files, self-improving stop/start hooks, path-scoped skills, an LSP-backed MCP server for symbol-level search, and subagents that split exploration from editing.
02
Explain the practical stakes without hype: New playlist item from Cole Medin; queued for transcript-backed review, topic mapping, and a practical learning artifact.
03
Map the idea onto the Inspect context -> Route tool -> Plan work -> Edit safely -> Verify behavior -> Report next step sequence and name the weakest link.
04
Produce the artifact and include the evidence that proves it: A coding-agent routing and execution matrix with context needed, tool choice, verification, and done signal.
Put it into practice
Give this grounded prompt to Codex or Claude after watching.
You are helping me turn one specific YouTube video into real, durable learning.
Source video:
- Title: Anthropic Just Dropped a Masterclass on Building Agent Harnesses (for Large Codebases)
- URL: https://www.youtube.com/watch?v=efRIrLXoOVA
- Topic: Agent Architecture
- My current learning frame: Take one real repo, add a lean root CLAUDE.md plus a subdirectory CLAUDE.md, wire up a self-improving stop hook, and install the plugin's LSP-backed search MCP, then run a scoped task to see the AI layer in action.
- Why this matters: New playlist item from Cole Medin; queued for transcript-backed review, topic mapping, and a practical learning artifact.
Transcript anchors from this exact video:
- 0:15 / Evidence 1: "already have a complex code base or two or three. You've got the apps and platforms that you're building, your second brain. You have these code bases that are tens or even hundreds of thousands of lines long..."
- 3:44 / Evidence 2: "And so, now with the AI layer, we have a third component of every code base introduced. This is everything like your global rules, your skills, your MCP servers and sub agents. Really every single individual feature of..."
- 15:35 / Evidence 3: "workflows and capabilities. And so, like this is an example of a skill right here for adding API routes in this code base. Really, a skill is some kind of set of steps, some kind of process, reusable..."
- 17:32 / Evidence 4: "claw.md. The distinction that I like to make is that global rules are your conventions. It's the rules that you need to follow. Like every route is registered here for example. Your skills are the workflows. So we..."
- 21:26 / Evidence 5: "kind of harness to give better search capability to Claude code when you're working in a larger code bases. And really they operate like skills. Just use sporadically throughout your session. So, like with skills we're loading in..."
- 23:31 / Evidence 6: "authentication." I don't know, I'm just kind of throwing out something off the cuff here, but you have sub agents built into a lot of these coding agents now, like Claude Code and CodeX. And so, you don't..."
- 27:33 / Evidence 7: "something that I help with. And so, I do offer enterprise trainings where I help you build out the AI layer, understand the core methodologies for AI coding, and create that standard for your adoption of coding agent..."
Video-aware target:
- Prompt lane: Coding-agent workflow
- Mechanism to extract: Find the workflow rule that explains when and how to use Codex, Claude Code, browser control, dashboards, or manual review.
- Artifact to produce: A coding-agent routing and execution matrix with context needed, tool choice, verification, and done signal.
- Artifact must include: task class; agent/tool choice; context packet; verification step; handoff/recovery rule
Your task:
1. Use the transcript anchors above as the primary source packet. If you add outside context, label it clearly as outside context and keep it secondary.
2. Create a source-check table with columns: timestamp, claim, transcript support, what the demo proves, confidence, and what still needs verification.
3. Extract the actual teachable mechanism from the video: Find the workflow rule that explains when and how to use Codex, Claude Code, browser control, dashboards, or manual review. Do not invent claims that are not supported by the title, lesson frame, or transcript anchors.
4. Build a reusable learning artifact: A coding-agent routing and execution matrix with context needed, tool choice, verification, and done signal.
5. Include:
- a plain-English definition of the core idea
- a diagram or structured model using this sequence: Inspect context -> Route tool -> Plan work -> Edit safely -> Verify behavior -> Report next step
- answers to these source questions: What workflow pain is being solved? | What exact coordination mechanism is shown? | What evidence proves it changes the work?
- 3 concrete examples that apply the video idea to real agentic work, such as rate-limit routing; browser verification after a UI edit; long-running /goal session review
- 2 failure modes the video helps prevent, chosen from the transcript evidence and these likely risks: choosing tools by hype; losing context across agents; letting parallel sessions become invisible
- a checklist for the next real workflow, focused on: routing decision, context portability, verification, handoff summary
- one practical exercise with a clear done signal: Route three recent tasks across Codex, Claude, browser checks, and manual review with a reason for each.
6. Add a "learning transfer" section: what changes in my workflow tomorrow if I actually learned this?
7. Add a "source check" section that cites which transcript anchor supports each major takeaway.
Quality bar:
- Make this specific to "Anthropic Just Dropped a Masterclass on Building Agent Harnesses (for Large Codebases)", not a generic Agent Architecture essay.
- Each workflow rule must point to a timestamped claim or demo moment, then state what remains unproven.
- Prefer operational examples, failure modes, and reusable artifacts over broad definitions.
- Call out uncertainty instead of smoothing over weak evidence.
- Avoid these generic drifts: generic Codex vs Claude comparison; feature lists without task routing; claims that ignore limits or recovery.
- If evidence is weak or missing, stop and say what transcript segment or timestamp needs review instead of guessing.
- Finish with a concise artifact I could paste into my learning app.
Misconceptions
What to stop believing.
A better model automatically makes a better agent.
The model matters, but harness design determines whether the system can act safely and repeatably.
More tools always help.
Every tool increases surface area. Strong agents have the right tools with clear permissions.
Memory means saving everything.
Useful memory is compressed, curated, and tied to future decisions.
Practice studio
Learning only counts when you make something.
01
Transcript evidence map
Separate what the video actually says from what you already believe about the topic.
3 source-backed takeaways with timestamps, confidence, and a transfer note.02
One useful artifact
Apply the video to a real workflow and produce a coding-agent routing and execution matrix with context needed, tool choice, verification, and done signal..
A reusable artifact with a done signal and one verification step.03
Coding-agent workflow teach-back card
Explain the coding-agent workflow mechanism to someone who has not watched the video yet.
A 90-second explanation, one diagram, one example, and one misconception to avoid.
Recall check
Answer first, then reveal — without rewatching.
How does Claude Code navigate a code base by default, and what is the trade-off?
What does the self-improving stop hook do at the end of a session?
Why use an LSP-backed MCP server instead of grep in very large code bases?
Source shelf
Use the video as a doorway, then verify with primary sources.
