I recently ran an overnight build of the swipe-study system, which included several interesting outputs. One standout was the Savannah Case Media SMM teardown, found at MarketingOS/docs/research/video/extracted/smm-agency/savannahcasemedia-smm-package-teardown.md. This teardown serves as a practical example of how we turn competitor videos into actionable research. The process operationalizes the Full-Funnel Swipe Doctrine, detailed in .claude/strategy/passive-brand-os/FULL-FUNNEL-SWIPE-DOCTRINE.md.
The swipe-study procedure begins with video transcription. We use a script located in E:\TE-Code\MarketingOS to transcribe videos. The command is simple: python scripts/transcribe-social-video.py <url> --theme <theme> --title "<slug>". If yt-dlp encounters issues, we reference TRANSCRIBE-PLAYBOOK.md to troubleshoot. This playbook includes steps like updating yt-dlp, using Firefox cookies, or capturing network traffic with a browser.
Once transcribed, verifying the author's identity is crucial. Many shared reel URLs point to collaborations or reposts, not the intended account. This discrepancy gets flagged to the user in our final summary, ensuring that our records are accurate.
Next, we capture the subject's online presence, starting with their website. We might use curl -sL or WebFetch, escalating to the site-capture skill if needed. We extract data such as follower counts, bio claims, and pinned content titles. These details, including package and pricing information, get recorded verbatim from screenshots.
Transcripts are then filed in MarketingOS/docs/research/video/transcripts/<theme>/ using a naming convention like YYYY-MM-DD-<account>-<topic>-<videoId>.md. These transcripts serve as immutable records, never to be edited post-commit.
The extracted teardown follows the structure in references/TEARDOWN-TEMPLATE.md. It covers various aspects such as the offer, delivery flow, and pricing. Importantly, any security flaws noted during capture—like exposed keys or unauthenticated steps—get added to client-factory/docs/core/SITE-SECURITY-BASELINE.md §1. This record helps us understand how competitors' weaknesses can inform our security practices.
The overnight build highlighted how these processes not only create detailed research artifacts but also contribute to our security baseline. By examining competitor strategies, we enhance our understanding and improve our own systems, ensuring that our operations remain robust and informed by the latest insights.
Get weekly insights on AI architecture, pattern recognition, and building platforms without permission.
The "speak-agent" skill is designed to make a web product usable by AI agents while proving its utility live. The overnight build system provided some...
Read itWhen you're tasked with giving a client's owner a dashboard on their own site, the owner-console skill in the Local Business OS is your go-to tool. This skill...
Read itResetting the overnight build system can be a daunting task, but it’s necessary to ensure everything is functioning as intended. I recently faced this during...
Read itHave thoughts on this post? I'd love to hear them! Join the conversation on X where we can discuss AI architecture, pattern recognition, and building platforms.
Discuss on XOr reach out directly at @TravisEric_