Commit graph

4 commits

Author SHA1 Message Date
f2330e68e0 Document the pr- campaign repository convention
Some checks failed
ci / build (push) Failing after 22s
Closes FLUID-WP-0009 T01. Writes down what a pr- repository is so the
second one need not be reverse-engineered from the first.

Keeps the workplan's insistence on naming propaganda alongside public
relations and promotion: the mechanism is identical in all three, and what
separates them is whether the campaign is accountable to something outside
itself -- a source it must not contradict, a named reviewer, a consent
basis it can lose.

Beyond the required contents, records the boundary table against fluid-*,
that a declared presence is brand and belongs to the campaign while the
credentials it produces belong to the interface, and that domain and topic
follow the subject being promoted rather than the activity -- the domains
are market sectors, and none of them is "marketing".

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_0172sgCZEEDJcnQmr4SGDvKa

Assistant: claude-code
Assistant-Model: opus
Assistant-Process: 1361245@bnt-lap001
Assistant-Session: b3b428ef-f3e6-4688-b091-01f71461d66a
2026-09-04 21:15:52 +02:00
4819fd919a Write SCOPE.md, and record that pr-hall-of-helix now exists
Some checks failed
ci / build (push) Has been cancelled
SCOPE.md was the register-generated stub. It now names the planes and the
packages behind them, the wire contract as the boundary, and what the
framework deliberately does not own -- the interfaces, the campaigns,
backend authority, and Phase D.

Adds three boundary tests, because most "does this belong in fluid-core?"
arguments reduce to one of them: could the control plane be killed right
now with no runtime effect; does this require the target API to change or
link anything; and would this be here if the Hall of Helix did not exist.

FLUID-WP-0009 T02 moves to progress. pr-hall-of-helix was created and
registered early, because the provisioning work needed a home for the
presence spec. The campaign itself has not moved, which is the part the
"after FT-WP-0001 T12" sequencing was protecting.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_0172sgCZEEDJcnQmr4SGDvKa

Assistant: claude-code
Assistant-Model: opus
Assistant-Process: 1361245@bnt-lap001
Assistant-Session: b3b428ef-f3e6-4688-b091-01f71461d66a
2026-09-04 21:11:35 +02:00
repo-manager
b40c0e01e1 repo.work.assign_missing_identifiers
Some checks failed
ci / build (push) Has been cancelled
source: repo-manager
reason: deterministic projection registration

Assistant: claude-code
Assistant-Model: opus
Assistant-Process: 1116572@bnt-lap001
Assistant-Session: 8ba9bb93-a72a-4883-b189-2499cce5c400
2026-09-04 11:59:31 +02:00
78454c59d8 Reframe the handover around composed posts, and plan the pr- pattern
The Telegram channel publishes posts written from hall entries, not the
entries themselves, so the handover's pressure, competition and
guardrails are rebuilt around that. Composition sits outside the
interface because Blueprint 48.1 forbids an LLM in the mandatory request
path, and because a post about a named person's work deserves a human
read that a request-path step would never get.

Adds FLUID-WP-0009 for the pr- campaign repository convention and for
reducing the Hall of Helix case from the thing this framework is for to
an example of using it. The prefix deliberately includes "propaganda":
a campaign repository that cannot name what it is doing will not govern
it well.

The workplan starts only after the loop has closed once on real traffic.
Extracting a seam before using it is how you learn you cut it in the
wrong place.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_014KmVxhJ35tCo7rE7UnLwWu

Assistant: claude-code
Assistant-Model: opus
Assistant-Process: 1116572@bnt-lap001
Assistant-Session: 8ba9bb93-a72a-4883-b189-2499cce5c400
2026-09-04 11:59:30 +02:00