--- schema: 1 type: video title: >- Skill issue: Lessons from skilling up coding agents to use Langfuse - Marc Klingen, Clickhouse description: >- Without a skill, Claude Code adds Langfuse using stale pre-training context, ships broken instrumentation, then catches the failure and fetches current docs to fix it. The resulting trace captures two LLM calls with no visibility into what the agent actually did. Marc Klingen covers the six learnings from building a skill to close that gap: surfacing a natural language search endpoint so agents stop crawling 478 documentation pages, why pointing to references beats duplicating content, and what happened when they ran an auto-research loop on the skill itself. Three of six suggested improvements shipped, but their target function nearly backfired by optimizing out the documentation-fetching steps that make the skill reliable over time. Speaker info: - https://x.com/marcklingen - https://www.linkedin.com/in/marcklingen/ Timestamps: 00:00 - Introduction to Marc Klingen and Langfuse 01:22 - Conceptual mental model for agent skills 04:04 - The problem: scaling documentation and onboarding for coding agents 09:03 - Six key learnings from building a Langfuse skill 09:12 - Learning 1: Looking at traces for debugging 10:47 - Learning 2: Helping agents navigate information (sitemaps/formats) 11:40 - Learning 3: Surfacing a search endpoint for docs 12:36 - Learning 4: Implementing basic evaluation (eval) setups 13:53 - Learning 5: Referencing rather than duplicating content 14:24 - Learning 6: Using auto-research loops with target functions 17:24 - Discussion on challenges: distribution, versioning, and target functions 19:21 - Roadmap and future outlook for agent automation 20:09 - Q&A session --- Watch along with the corrected transcript, summary, timestamps and resources on this talk's official AIE page: https://ai.engineer/talks/vNCY9kXXyDQ-skill-issue-lessons-from-skilling-up-coding-agents-to-use-langfuse Join our upcoming AI Engineer conferences (see https://ai.engineer landing page): - AIEi Shanghai, November 5–6 — the first for meeting top Chinese labs and engineers - AIE CODE in SF, November 10–12 — the top AI Coding conference returns - AIEi Sydney, December 7–8 — the first AIEi run alongside NeurIPS 2026 in Sydney --- canonical: 'https://liquidgold.cc/watch/vNCY9kXXyDQ' markdown: 'https://liquidgold.cc/watch/vNCY9kXXyDQ.md' llms: 'https://liquidgold.cc/watch/vNCY9kXXyDQ/llms.txt' parent: 'https://liquidgold.cc/channel/aidotengineer.md' channel: aidotengineer updated: '2026-10-24T13:28:41.145Z' language: null children: [] videoId: vNCY9kXXyDQ durationSeconds: 1449 publishedAt: '2026-05-20T15:00:06.000Z' searchChunks: - ordinal: 0 text: null startSeconds: 25 endSeconds: 113 watch: 'https://liquidgold.cc/watch/vNCY9kXXyDQ?t=25' heldReason: rights - ordinal: 1 text: null startSeconds: 25 endSeconds: 113 watch: 'https://liquidgold.cc/watch/vNCY9kXXyDQ?t=25' heldReason: rights - ordinal: 2 text: null startSeconds: 123 endSeconds: 244 watch: 'https://liquidgold.cc/watch/vNCY9kXXyDQ?t=123' heldReason: rights - ordinal: 3 text: null startSeconds: 244 endSeconds: 255 watch: 'https://liquidgold.cc/watch/vNCY9kXXyDQ?t=244' heldReason: rights - ordinal: 4 text: null startSeconds: 255 endSeconds: 353 watch: 'https://liquidgold.cc/watch/vNCY9kXXyDQ?t=255' heldReason: rights - ordinal: 5 text: null startSeconds: 365 endSeconds: 465 watch: 'https://liquidgold.cc/watch/vNCY9kXXyDQ?t=365' heldReason: rights - ordinal: 6 text: null startSeconds: 365 endSeconds: 465 watch: 'https://liquidgold.cc/watch/vNCY9kXXyDQ?t=365' heldReason: rights - ordinal: 7 text: null startSeconds: 476 endSeconds: 568 watch: 'https://liquidgold.cc/watch/vNCY9kXXyDQ?t=476' heldReason: rights - ordinal: 8 text: null startSeconds: 476 endSeconds: 568 watch: 'https://liquidgold.cc/watch/vNCY9kXXyDQ?t=476' heldReason: rights - ordinal: 9 text: null startSeconds: 579 endSeconds: 667 watch: 'https://liquidgold.cc/watch/vNCY9kXXyDQ?t=579' heldReason: rights - ordinal: 10 text: null startSeconds: 579 endSeconds: 667 watch: 'https://liquidgold.cc/watch/vNCY9kXXyDQ?t=579' heldReason: rights - ordinal: 11 text: null startSeconds: 667 endSeconds: 678 watch: 'https://liquidgold.cc/watch/vNCY9kXXyDQ?t=667' heldReason: rights - ordinal: 12 text: null startSeconds: 678 endSeconds: 776 watch: 'https://liquidgold.cc/watch/vNCY9kXXyDQ?t=678' heldReason: rights - ordinal: 13 text: null startSeconds: 678 endSeconds: 776 watch: 'https://liquidgold.cc/watch/vNCY9kXXyDQ?t=678' heldReason: rights - ordinal: 14 text: null startSeconds: 678 endSeconds: 776 watch: 'https://liquidgold.cc/watch/vNCY9kXXyDQ?t=678' heldReason: rights - ordinal: 15 text: null startSeconds: 787 endSeconds: 896 watch: 'https://liquidgold.cc/watch/vNCY9kXXyDQ?t=787' heldReason: rights - ordinal: 16 text: null startSeconds: 787 endSeconds: 896 watch: 'https://liquidgold.cc/watch/vNCY9kXXyDQ?t=787' heldReason: rights - ordinal: 17 text: null startSeconds: 787 endSeconds: 896 watch: 'https://liquidgold.cc/watch/vNCY9kXXyDQ?t=787' heldReason: rights - ordinal: 18 text: null startSeconds: 896 endSeconds: 907 watch: 'https://liquidgold.cc/watch/vNCY9kXXyDQ?t=896' heldReason: rights - ordinal: 19 text: null startSeconds: 907 endSeconds: 1017 watch: 'https://liquidgold.cc/watch/vNCY9kXXyDQ?t=907' heldReason: rights - ordinal: 20 text: null startSeconds: 907 endSeconds: 1017 watch: 'https://liquidgold.cc/watch/vNCY9kXXyDQ?t=907' heldReason: rights - ordinal: 21 text: null startSeconds: 1028 endSeconds: 1131 watch: 'https://liquidgold.cc/watch/vNCY9kXXyDQ?t=1028' heldReason: rights - ordinal: 22 text: null startSeconds: 1028 endSeconds: 1131 watch: 'https://liquidgold.cc/watch/vNCY9kXXyDQ?t=1028' heldReason: rights - ordinal: 23 text: null startSeconds: 1290 endSeconds: 1449 watch: 'https://liquidgold.cc/watch/vNCY9kXXyDQ?t=1290' heldReason: rights --- # Skill issue: Lessons from skilling up coding agents to use Langfuse - Marc Klingen, Clickhouse Without a skill, Claude Code adds Langfuse using stale pre-training context, ships broken instrumentation, then catches the failure and fetches current docs to fix it. The resulting trace captures two LLM calls with no visibility into what the agent actually did. Marc Klingen covers the six learnings from building a skill to close that gap: surfacing a natural language search endpoint so agents stop crawling 478 documentation pages, why pointing to references beats duplicating content, and what happened when they ran an auto-research loop on the skill itself. Three of six suggested improvements shipped, but their target function nearly backfired by optimizing out the documentation-fetching steps that make the skill reliable over time. Speaker info: - https://x.com/marcklingen - https://www.linkedin.com/in/marcklingen/ Timestamps: 00:00 - Introduction to Marc Klingen and Langfuse 01:22 - Conceptual mental model for agent skills 04:04 - The problem: scaling documentation and onboarding for coding agents 09:03 - Six key learnings from building a Langfuse skill 09:12 - Learning 1: Looking at traces for debugging 10:47 - Learning 2: Helping agents navigate information (sitemaps/formats) 11:40 - Learning 3: Surfacing a search endpoint for docs 12:36 - Learning 4: Implementing basic evaluation (eval) setups 13:53 - Learning 5: Referencing rather than duplicating content 14:24 - Learning 6: Using auto-research loops with target functions 17:24 - Discussion on challenges: distribution, versioning, and target functions 19:21 - Roadmap and future outlook for agent automation 20:09 - Q\&A session --- Watch along with the corrected transcript, summary, timestamps and resources on this talk's official AIE page: https://ai.engineer/talks/vNCY9kXXyDQ-skill-issue-lessons-from-skilling-up-coding-agents-to-use-langfuse Join our upcoming AI Engineer conferences (see https://ai.engineer landing page): - AIEi Shanghai, November 5–6 — the first for meeting top Chinese labs and engineers - AIE CODE in SF, November 10–12 — the top AI Coding conference returns - AIEi Sydney, December 7–8 — the first AIEi run alongside NeurIPS 2026 in Sydney --- Canonical: https://liquidgold.cc/watch/vNCY9kXXyDQ · Channel: [AI Engineer](https://liquidgold.cc/channel/aidotengineer.md) > Search chunk text is held until public-source provenance is verified. ## Document chunks > Missing: no stored document chunks are available.