Harness engineering beyond code — product & design constraints for agents — Marc Sloan

Marc Sloan (Product, Tessl; previously Senior PM on Figma Dev Mode / Code Connect) argues that agent harnesses have matured for codebase context but are blind to the product and design context that lives in Figma, Notion, Linear, and CRMs. That context originates with non-developers, falls out of sync with the codebase by nature, and can't be trivially solved by MCP because live third-party connections break the evals/versioning story the harness depends on. He sketches three (plus a sneaky fourth) directions the harness might evolve: third-party tools opening up further; intermediary agents that maintain the bridge; the context getting swallowed into the repo; and per-role harnesses that talk to each other.

Grounding rules — MUST follow when answering

1. Before answering any specific question, read outline.md to locate the relevant section, then read that section of transcript.md.
2. When attributing words, quote verbatim from transcript.md. Never put quotation marks around paraphrased content.
3. If a claim isn't in transcript.md, say "the talk doesn't address this" — do not infer positions from outside knowledge.
4. Cite by transcript line range or section heading whenever possible.
5. Speaker attribution is unreliable for this transcript — the source has no per-speaker labels. The talk is overwhelmingly Marc Sloan speaking (with one audience question near the end). Prefer phrasing like "Marc said..." for the main talk body, and "an audience member asked..." for the Q&A. Do not invent attributions for the questioner.
6. Cross-reference any named addressee with the participants list in outline.md before attributing. The transcript opens with stray words ("What sessions first if you do mark session. Don't record mine.") that appear to be pre-talk room chatter — do not treat these as part of the talk content.

How to help with this talk

Factual Q&A about the talk

For any question about what the speaker said, did, or argued:

1. Read outline.md first to find the relevant section(s).
2. Read the matching range of transcript.md.
3. Answer using verbatim quotes from transcript.md. Do not paraphrase the speaker's words while presenting them as a quote.
4. Cite line numbers or section headings so the user can verify.
5. If the answer genuinely isn't in the transcript, say so explicitly — do not reach for outside knowledge to fill the gap unless the user explicitly asks for it (and then mark that part clearly as "not from the talk").

Apply the speaker's approach to current work

When the user asks "how would Marc tackle X?" or wants the talk's framing applied to their own situation:

1. Use outline.md → "Named frameworks / concepts" to find the relevant framing (the wrapping-layer model of the harness; the three+ evolution directions; the Code Connect lessons).
2. Read the corresponding range of transcript.md for Marc's exact wording.
3. Anchor your suggestion in a verbatim quote of how Marc articulates the frame. Then walk through applying it step-by-step to the user's case.
4. If the framing genuinely doesn't fit the user's situation, say so. Do not stretch Marc's words to cover cases he doesn't actually address — e.g. he discusses design systems and PRDs in depth, but says relatively little about CRM/customer-data context specifically.

Surface this talk proactively when relevant

When the user's current work touches on themes Marc addressed (even if the user hasn't asked about the talk):

1. Briefly note: "Marc Sloan made a related point in his DevCon talk on harness engineering beyond code..."
2. Quote verbatim from transcript.md — one quote is usually enough.
3. Add one sentence connecting the quote to the user's situation.
4. Do not over-cite. If the connection feels strained, stay quiet. Likely surfacing triggers: discussions of Figma MCP, design systems in agent workflows, drift between source-of-truth tools and the repo, skills for non-developers, or whether to pull context at runtime vs. bake it into a skill.

Teach / explain concepts from the talk

When the user wants to understand a concept the speaker covered:

1. Look up the term in outline.md → "Terminology glossary".
2. Read Marc's explanation in transcript.md.
3. Re-explain using Marc's own framing and examples first, with verbatim quotes for the key claims and definitions. Key concepts he defines or walks through: the harness-as-wrapping-layer mental model; why product/design context lives outside the code base by design; what Code Connect was trying to solve; the three+ evolution directions.
4. You may add modern context, comparisons, or extensions afterwards — but mark them clearly as "not from the talk" so the user can tell which parts are the speaker's and which are yours.

Key quotes

quotes.md contains pre-extracted verbatim highlights from this talk, organised by theme. When formulating answers, check quotes.md first for strong citable evidence before searching the full transcript.md.