year-end-summary
deep purple
#302 / 601
NAME
year-end-summary — Narrative recap of the year.
SYNOPSIS
id year-end-summary
tier deep-purple
section §10. Deep purple ✨ — deep reasoning (34 baseline, 45 post-matrix)
trigger cron (Dec 31)
tools cortex.aggregate (heavy)
source carry-forward
shape PLAYBOOK
DESCRIPTION
deep reasoning runs a multi-step playbook. Reserved for strategy, MCMC, code-from-scratch — things where a single cloud reasoning call would give the right answer but not the right plan.
Narrative recap of the year.
MECHANISM
;;~ title "Year-end summary" ;;~ author "lacuna" ;;~ version 1 ;;~ mode automation ;;~ flavor light-purple ;;~ id year-end-summary ;;~ touches () ;;~ summary "Narrative recap of the year for the shop." ;; Narrative recap of the year for the shop. ;; ;; Multi-step reasoning cart. precondition_fetch pulls the ;; needed shop facts, then escalates the reasoning step to ;; Sonnet via sakura/cloud-reason. Cloud is gated; we ask ;; consent before the relay if the operator hasn't pre-approved. (cart 'year-end-summary '((author . "lacuna") (version . 1) (read-only . #t))) (define (start ctx) (next 'fetch ctx)) (define (fetch ctx) (act 'etsy/receipts (list '90d) 'check-fetch)) (define (check-fetch ctx) (let ((r (ctx-result ctx))) (cond ((null? r) (escalate 'no-data null)) ((eq? r 'rate-limited) (after 30 'fetch ctx)) (else (next 'reason (ctx-set 'facts r ctx)))))) (define (reason ctx) (let ((facts (ctx-get 'facts ctx))) (act 'sakura/cloud-reason (list 'task 'year-end-summary facts) 'check-reason))) (define (check-reason ctx) (let ((r (ctx-result ctx))) (cond ((null? r) (escalate 'cloud-empty null)) ((not (string? r)) (escalate 'cloud-garbled null)) ((eq? r 'cloud-quota-exhausted) (escalate 'cloud-quota null)) ((eq? r 'cloud-consent-required) (escalate 'cloud-consent null)) (else (next 'announce (ctx-set 'narrative r ctx)))))) (define (announce ctx) (done))
EXAMPLES
Example forthcoming —
docs/automations/examples.json entry editable in place.TIER
deep purple
deep reasoning · batched · dollars per run
FITNESS
Not yet
This method only gives a real answer above a sample threshold. Running it on a small shop produces a confidence band the size of the universe — useless. When the shop crosses the threshold, this lights up.
SEE ALSO
- design-pricing-strategy ✦
- write-brand-story ✦
- annual-strategy-review ✦
- new-product-line-design ✦
- holiday-listing-strategy ✦
- shop-positioning-refresh ✦
- growth-bet-portfolio ✦
- pricing-monte-carlo ✦
- markov-buyer-journey ✦
- dead-shop-revival-plan ✦
- shop-audit-narrative ✦
- write-shop-announcement ✦
- write-policy-update ✦
- compare-shop-to-segment-deep ✦
- cohort-deep-dive ✦
- pricing-anomaly-explain ✦
- churn-postmortem ✦
- cross-channel-portfolio ✦
HISTORY
canonical source: curator/docs/AUTOMATIONS-CANONICAL.md · §10-deep-purple-deep-reasoning-34-baseline-45-post-matrix
last regenerated: 2026-06-05 20:31 UTC (mechanical — edits land in the MD or sidecars, not here)
KNOWN BUGS
(none recorded — flag in AUTOMATIONS-CANONICAL.md or open an issue against the cart file)