New workflow rule: grep daily-insights for hard data before creating working repos
Source type:
obs· Harvested: 2026-05-03 · Original date: 2026-05-03T11:19:04.582Z Metadata:{"project":"lunhsiangyuan","type":"decision","obs_id":64966}
obs/64966 · decision · 2026-05-03T11:19:04.582Z
New workflow rule: grep daily-insights for hard data before creating working repos
A new workflow rule was established to prevent duplicate research when creating working repositories. Before building any new project repo (like garnet-negotiation, cms-access-model, or e2-immigration), there must be a systematic grep of daily-insights for hard data—specifically dollar amounts, percentages, and entity identifiers using regex patterns. The rule was created after discovering that Abourbih’s 716K compensation figures had already been researched and logged in April 10 daily-insights, but were missed when initially building the garnet-negotiation repo. This led to using MGMA estimates instead of actual Form 990 data, requiring retroactive correction of the entire walk-away table. The workflow specifies a two-step grep: first for keywords to find relevant files, then for hard data patterns within those files, examining at least 200-300 lines rather than stopping early.
Concepts: [“pattern”,“gotcha”,“why-it-exists”]
Facts: [“New feedback rule requires grepping daily-insights for hard data patterns before creating any working repo”,“Hard data includes dollar amounts (XXXK), percentages, and EINs using specific regex patterns”,“Rule triggered by wasting research effort re-discovering Garnet Form 990 data that was already in 2026-04-10 daily-insights”,“Cross-check must examine 200-300 lines, not just head -120, to catch all relevant data”]
[← 回 Alfred Brain Hub]