jiang-episode-read-writerlisted

# Jiang Episode Read Writer Use this after a source transcript and semantic outputs exist. The output is the public episode read: ```text content/lens/episodes/<source-slug>/read.json ``` This skill is editorial. It does not import Colab files, process transcript packets, rewrite lens docs, or promote canon. ## Inputs - `content/sources/videos/<source-slug>/transcripts/v1/transcript.clean.jsonl` - `content/lens/evidence/videos/<source-slug>.semantic.json` - existing strong reads, especially `content/lens/episodes/predictive-history-0hyet47cc-e/read.json` - existing lens point IDs from compiled link index when adding lens links ## Reading Pass Before writing: 1. Read the full clean transcript, not only extracted claims. 2. Inspect semantic outputs for signature moments, real questions, quoted material, speaker uncertainty, and chronology notes. 3. Identify the episode's argumentative movement in one sentence. 4. Identify the sharpest Jiang formulations that must survive compression. 5. Check whether the episode directly invokes existing lens points. ## Corpus Anchor Pass Before finalizing the read, compare the source against the existing corpus: - strong prior episode reads, especially ones already curated by maintainers, - existing lens pages and lens-point anchors, - topic clusters and aliases, - semantic `signature_moments` from the current source, - prior Jiang phrases that sound close to this source. Use the comparison to avoid two failures: a bland summary th