From thinking-frameworks-skills
Appends structured YAML observation blocks to audience-notes.md for weekly pipeline insights with confidence ≥ medium. Append-only, evidence-backed, curator-reviewed.
How this skill is triggered — by the user, by Claude, or both
Slash command
/thinking-frameworks-skills:update-audience-notesThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
```
At end of weekly pipeline:
- [ ] Step 1: Review attribution outputs for any pattern observed
- [ ] Step 2: Filter to confidence ≥ medium only (low is never appended)
- [ ] Step 3: Format as YAML block per schema
- [ ] Step 4: Append to shared-context/audience-notes.md
- [ ] Step 5: If no medium+ observation this week, log "no additions this week" to weekly report's data-caveats (not to audience-notes)
- date: YYYY-MM-DD
week: YYYY-WW
observation: >
{plain English, ideally testable}
confidence: medium | high
evidence:
- slug: {post-slug}
metric: {value}
baseline: {value}
reviewed_by_curator: false
- date: 2026-06-01
week: 2026-W22
observation: >
External restack from a larger ML account on Substack Notes is
the strongest sub-acquisition channel observed to date —
+49 subs from a single post vs ~+4 baseline, with 71% of the
new cohort landing in "active" tier immediately.
confidence: high
evidence:
- slug: why-your-embedding-search-melted-at-10pm
open_rate: 0.61
click_rate: 0.12
baseline_open: 0.49
baseline_click: 0.05
new_subs: 49
active_tier_of_new_subs: 35
external_restacks_observed: 2
reviewed_by_curator: false
reviewed_by_curator: false on append; Curator flips to true on its next monthly review.observation field's content.npx claudepluginhub lyndonkl/claude --plugin thinking-frameworks-skillsComposes a structured weekly growth report (400-800 words) from upstream analytics outputs, enforcing YAML frontmatter schema, seven-section body, and word budget with weakest-section truncation.
Run the daily content pipeline to fetch signals, analyze relevance, draft output, edit for voice fidelity, and deliver the brief
Generates Substack Note ideas by scanning YouTube videos, newsletters, and prior Notes. Orchestrates fetching, processed-log management, duplicate prevention, and delegation to idea extraction. Use for content repurposing and posting cadence.