dhkim2750's honest review, Claude Code Harness & Graph Engineering: Advanced Claude Code CLI Harness Engineering in Practice course
Reviews 1
Average rating 3
Before writing this course review, I would like to mention that I am already a developer with a certain level of prior knowledge, and that I enrolled in this course with the intention of using what I learned immediately in my professional work. I believe the instructor’s expertise and the level of the course content are both extremely high, and that this is a truly excellent course for learning the overall theory behind agents. Since the only way to properly understand why they work and the logic behind them is to read the relevant papers thoroughly, this course was very helpful in that regard. However, from my perspective as someone who enrolled in the course to apply the material directly to my work, I found it somewhat disappointing that learners had to make the connection between the theoretical content and its practical application themselves. I also found myself wondering whether all of the theoretical material was really necessary to devote an entire session to. That said, viewed as a course for studying the subject academically, I think it is a good course. (However, like many others, I agree that the provided course materials are quite lacking.) I would also like to create and open-source well-known skills such as ponytail myself, and I think this course will be very helpful in building a solid theoretical foundation.
1
Hello Dahyun Kim, Thank you so much for taking the time to leave such a thoughtful review. As I read your feedback, it became clear which areas I need to improve when creating the course. In particular, understanding the theory itself and connecting that theory to designing Agents or implementing Skills in real-world work are two different things. It seems that, in the current course, learners had to make too many of those connections themselves. I want to continue providing an in-depth explanation of why Agents work the way they do, but going forward, I plan to improve the course so that it also shows where each theory is used in actual Agent implementations, what problems it solves, and how it connects to code and Skills. For example, rather than simply ending with explanations of papers and concepts, I will review examples and exercises that connect the following steps: Theory → Actual Agent Architecture → Implementation → Failure Cases → Practical Application Guidelines I also think that directly designing and implementing a real Skill like the Ponytail you mentioned would be a direction that fits the course well. I will prepare to add exercises in which learners design, implement, and validate a single feature from start to finish. The course materials on the GitHub repository were intended to allow learners to review and practice both the theory and the exercises. Were they still insufficient? Thank you again for appreciating the theoretical depth of the course. I will preserve that strength while improving the course so that, after taking it, learners can naturally progress to the question, “So how do I actually build this?” Thank you very much for your valuable feedback.
0