Skip to content

social/media cloud transfer: blog desk, reels law, and promotion kernel - #13532

Merged
cursor[bot] merged 20 commits into
masterfrom
cursor/social-media-cloud-transfer-462b
Nov 9, 2026
Merged

cursor[bot] merged 20 commits into
masterfrom
cursor/social-media-cloud-transfer-462b

Conversation

@gHashTag

@gHashTag gHashTag commented Nov 9, 2026

Copy link
Copy Markdown
Collaborator

Description

Moves the open social and media work that can be finished in this repository onto one cloud branch. Nothing was posted to X, Instagram, Telegram, Facebook, Threads, or LinkedIn. No post was scheduled. BLOG_AUTOPOST ships at 0.

Pull Request Checklist

  • Specs changed are .t27 with tests. English and ASCII only.
  • t27c test-report on every touched spec: FAIL 0, vacuous 0.
  • Seals were not re-saved. This environment has a debug t27c (build 0 commit unknown), not the master release binary. A seal from it would record the wrong toolchain.

Finished here

Closes is on the commit only when that commit finishes the issue.

Advanced, issue left open

Testing

./target/debug/t27c test-report specs/automation/blog-share.t27
./target/debug/t27c test-report specs/automation/meta-connect.t27
./target/debug/t27c test-report specs/automation/promotion_kernel.t27
./target/debug/t27c test-report specs/tri/blog/blog.t27
./target/debug/t27c test-report specs/blog/epic_progress.t27
./target/debug/t27c test-report specs/blog/t27b_progress.t27

Each report printed FAIL 0 and vacuous passes 0.

Documentation

docs/README.md indexes docs/promotion/. The cards point at specs/automation/promotion_kernel.t27. If a number disagrees, the spec wins.

Review notes

No new foreign file. No hand-written Rust, Python, or workflow. Duplicate-body scan on the touched specs reports nothing written elsewhere. Seals for the edited specs are now stale relative to these bytes and should be re-saved with a master-built t27c.

Open in Web Open in Cursor 

cursoragent and others added 20 commits November 9, 2026 18:17
The desk could ask about one article a day and had no Telegram receipt
and no backlog slot. v6 states the new-article cap, the one-use UTC
slot, queue-then-newest order, and the Telegram message_id rule. The
shipped switch stays 0, so this commit does not post or schedule.

Co-authored-by: Dmitrii Vasilev <gHashTag@users.noreply.github.com>
An image container cannot carry a finished reel. v3 names REELS,
the one-minute poll, the 2026-10-09 limits, and the rule that a
media_publish call is not repeated. A video still needs a yes, and
the host file is not in this repo, so the issue stays open.

Co-authored-by: Dmitrii Vasilev <gHashTag@users.noreply.github.com>
#13466)

The live catalog is 109 records and 13 clusters. A draft that says
84 formats, BF16 21/21, FP8 15/16, or the paper's 12 clusters is
refused. The contract does not post.

Co-authored-by: Dmitrii Vasilev <gHashTag@users.noreply.github.com>
…13467)

A fitting claim with no visual stays a draft. asset_ok asks for a
claim, an artifact, a path, a visual, a CTA and a metric. may_post
stays false, so readiness does not publish.

Co-authored-by: Dmitrii Vasilev <gHashTag@users.noreply.github.com>
day_claim names a claim for each of the ten days. BF16 21/21, FP8
15/16 and 84 formats stay refused. A visual does not make a day
publishable: publishable_days is 0.

Co-authored-by: Dmitrii Vasilev <gHashTag@users.noreply.github.com>
metric_known is true for the 7-day impression count cited by #13465
and false for clicks, course starts and payments. Those counts are
not stored as zero. Paid checkout stays off.

Co-authored-by: Dmitrii Vasilev <gHashTag@users.noreply.github.com>
An unchecked claim, a missing visual, or paid-token use stops the
lane. repair_hides_failure stays false, so a repair cannot clear
the stop by hiding it.

Co-authored-by: Dmitrii Vasilev <gHashTag@users.noreply.github.com>
pk_event builds a t27l/v1 line on surface agent. tl_validate_line
accepts it. The payload is the claim id claim-109, and a payload
that carries an address is refused.

Co-authored-by: Dmitrii Vasilev <gHashTag@users.noreply.github.com>
…13477)

Each calendar day has a pillar tag and a claim id. The week reviews
those drafts and weekly_posts stays 0. The markdown card points at
the spec so the cycle has one executable home.

Co-authored-by: Dmitrii Vasilev <gHashTag@users.noreply.github.com>
The bank holds the 109-record catalog and the 13-cluster gap. Both
are claims the gate marks fit. 84 formats is not a draft. Neither
draft has a visual, and neither is published.

Co-authored-by: Dmitrii Vasilev <gHashTag@users.noreply.github.com>
publish_gate is true only when a widget, a screenshot, a passing
test and a human yes are all present. gate_transmits stays false,
so the gate does not send a post.

Co-authored-by: Dmitrii Vasilev <gHashTag@users.noreply.github.com>
…13474)

Stars, forks, open issues and contributors were read from the GitHub
API on 2026-11-10. X impressions stay the 2026-11-09 citation in
#13465 because this agent did not sign in. Course readiness is not
measured. Nothing was posted.

Co-authored-by: Dmitrii Vasilev <gHashTag@users.noreply.github.com>
The sixteen field names match the epic final-report template. Only
the 7-day impression count is a known metric. Clicks, course starts
and payments stay blank rather than zero.

Co-authored-by: Dmitrii Vasilev <gHashTag@users.noreply.github.com>
The funnel is eight steps from X through B2B. Each step has one CTA.
Paid product and B2B are named and not offered: paid checkout stays off.

Co-authored-by: Dmitrii Vasilev <gHashTag@users.noreply.github.com>
plan, draft, check and publish are specified with tests, including
70/149 rounding to 47 and a refused publish without --yes. The law
never posts to a social network. The cli/tri glue is a new foreign
file and is not in this commit, so the issue stays open.

Co-authored-by: Dmitrii Vasilev <gHashTag@users.noreply.github.com>
The existing file is one post, not the card law. This adds the
1200x675 layout, the not-measured rule, and when a post is due
(no previous card, a 5-point move, or a milestone change). The
37 draft posts and the CLI glue are not in this commit.

Co-authored-by: Dmitrii Vasilev <gHashTag@users.noreply.github.com>
The post stopped at 380 checked of 723 on 2026-10-05. The lab summary
named in the issue is 2454 checked of 2602, 4 vacuous, 144 blocked,
on master 5e47682. The headline and the shown percents follow the
spec's own rounding. This commit does not publish the post to t27.ai.

Co-authored-by: Dmitrii Vasilev <gHashTag@users.noreply.github.com>
card_points was written twice, which grows the duplicate-body
ratchet. The card spec now calls the one function in
specs/tri/blog/blog.t27. The not-measured sentinel stays local
because an imported const does not survive inlining.

Co-authored-by: Dmitrii Vasilev <gHashTag@users.noreply.github.com>
The new docs/README heading used an em dash. First-party markdown
stays ASCII, so the heading uses a hyphen.

Co-authored-by: Dmitrii Vasilev <gHashTag@users.noreply.github.com>
The GitHub read was 2026-11-09 UTC, which is 2026-11-10 in
Asia/Bangkok. BASELINE_DATE stays that Bangkok day. The card no
longer calls the day UTC.

Co-authored-by: Dmitrii Vasilev <gHashTag@users.noreply.github.com>
@github-actions

github-actions Bot commented Nov 9, 2026

Copy link
Copy Markdown
Contributor

PR Dashboard

Generated at: 2026-11-09 18:34:04 UTC

Summary

Status Count
Total Open PRs 50
PRs with Failing Checks 42
PRs with All Checks Green 8
READY 3
FAILING 42
PENDING 0
NO CHECKS YET 0

These columns do not partition: 3 + 42 + 0 + 0 = 45, and there are 50 open PRs. A PR is being counted twice or not at all.

Seal Status

  • ⚠️ STALE -- sha256(compiler.rs)=b882be1e4933 != manifest seal=87e5cbd3ad94.
    The committed NMSE numbers were certified against an older compiler.rs.
    Run scripts/reseal-check.sh locally for the two-step reseal command (advisory; not a merge gate).

@cursor
cursor Bot merged commit 39dd576 into master Nov 9, 2026
33 of 38 checks passed
@cursor
cursor Bot deleted the cursor/social-media-cloud-transfer-462b branch November 9, 2026 19:56
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment