Skip to content

Registry 2026.11.03-registry.40 on Core.51: Telegram Content approves the post itself; llm-call/openrouter accepted - #70

Merged
0design merged 11 commits into
mainfrom
claude/registry40-core51-1103
Nov 3, 2026
Merged

0design merged 11 commits into
mainfrom
claude/registry40-core51-1103

Conversation

@0design

@0design 0design commented Nov 3, 2026

Copy link
Copy Markdown
Owner

Registry and tests only; the Core runtime does not change (packed archive stays 59294943…2f4 = Core.51 on main).

Includes the seven commits of #69 (every sending Registry workflow holds at a human approval before it sends, and sends what was approved). On top:

  • Telegram Content (content-factory 1.3.0): «Publish this?» binds with sha256, and the Telegram message text comes from the approved subject. On Core.51 (Gates: human gate approves the content past agent verdicts; agent-gate budget from maxTokens, cut-off answers escalate #63) the human gate's subject is the post the agent gate judged; the machine verdict is listed in reviews, not in the subject or its hash. Tests: the gate subject is the post, the verdict is only in reviews, the hash does not change with the verdict, the body after approval equals the approved text, nothing is sent before the decision.
  • llm-call/openrouter 1.2.0 accepted: live run on the installed Core.51 archive with a Keychain key (2 bounded calls, provider-reported cost 0.003436 USD; an unpriced model is refused before any call). Evidence registry/demos/llm-call-openrouter.txt; digest is now components-ready.
  • releaseVersion 2026.11.03-registry.40-candidate. The site pins it for stage only; nothing is published and no production change.

Supersedes #69.

🤖 Generated with Claude Code

0design and others added 11 commits November 1, 2026 21:29
… and do what they say

Candidate 2026.11.01-registry.39 on the Core.49 base (6814eb7). Only registry/, test/ and
scripts/ change, so `npm pack --ignore-scripts` still gives the pinned Core archive
631ba858051dcf08b4dbd204d12c50ed07c5b1901fadf5cdc8d9043905778394.

- content-feed, feed-fanout, release-watch, webhook-relay, json-digest, brand-mentions: a human
  approval-gate (bind: sha256) directly before the send; the send carries the gate's subject,
  so what is sent is what was approved. feed-fanout gates each lane.
- json-digest and webhook-relay send the approved object in Core's envelope: titles and data
  stay JSON instead of text inside a string.
- brand-mentions: the brand is the required input `brand`; the schedule trigger is removed
  (a scheduled run has no inputs and would fail every day).
- price-watch: escalateOn none, so an ordinary day ends at the check without asking anyone;
  the reason states the rate; the alert sends the approved reason.
- content-factory: model set in the manifest. The approval subject is still the check's
  verdict on Core.49 (needs the next Core release); the brief now says so.
- Every changed workflow sets settings.model (no unconfigured model); versions 1.1.0 -> 1.2.0.
- test/registry-real-workflows.test.mjs: offline runs (stubbed fetch) prove nothing is sent
  before the gate and the post-approval body equals the approved subject; all 11 fail on the
  registry.38 bytes. core37/registry-build/export tests follow webhook-relay now being gated.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…man gate approves the content past the verdict)

On Core.51 a human gate's subject is the content past agent verdicts, so the alert reads the
rate from the approved subject. The Core.49 stage candidate (branch -core49) keeps the
verdict reason, which is its subject there. content-factory's brief drops the Core.49 note.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…nd webhook-relay (the visual editor's insert keeps them in place)

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…ence) once, then sends each entry from it

On Core.49 lane rows share step ids, so a gate inside the second lane took the previous
lane's delivery as its subject and its sentence never reached the receiver; the title and
link were sent without being shown either. One model call now returns the list, one human
gate approves it exactly, and the fan-out runs over the approved list. price-watch's brief
names the no-key case.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…n words in the editor)

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…t an engine guarantee); price-watch's no-key limit is true for this Core

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…ate component

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Release registry.40 candidate with content-factory 1.3.0 and reproducible Core.51 checksums.

Co-Authored-By: Codex <noreply@openai.com>
…its verdict

Check the human stop, subject hash, approved send body and independence from review reasons offline.

Co-Authored-By: Codex <noreply@openai.com>
…ve installed-archive run

A live run on the packed Core.51 archive (59294943…2f4, installed with npm, key from the macOS
Keychain): two bounded OpenRouter calls with an explicit model, retries/timeout, a per-call
money cap and provider-reported cost (0.003436 USD); an unpriced model is refused before any
call. Evidence: registry/demos/llm-call-openrouter.txt. Implementation implemented, acceptance
accepted with the evidence bound to componentVersion 1.2.0 and the Core archive; digest is now
components-ready. Catalog, composition and SHA256SUMS rebuilt with the reproducible archive.

Co-Authored-By: Codex <noreply@openai.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…records show (no unlinked billing line)

Review r1 (P2): the billing check was a separate call, not linked to the two acceptance calls.
Evidence re-hashed (eacfb9bb…); catalog and SHA256SUMS rebuilt with the reproducible Core archive.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@0design
0design merged commit 7706044 into main Nov 3, 2026
2 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant