fluid-core/workplans/FLUID-WP-0008-fluid-telegram-handover.md
tegwick 3c4eb1125d
Some checks failed
ci / build (push) Failing after 3h10m20s
Close FLUID-WP-0009 T03 and mark blocked workplans as such
Audited fluid-core for accidental coupling to the Hall of Helix campaign
(redaction defaults, classifier thresholds, fitness metric names,
examples/echo-interface): none found beyond the deliberate conformance
handover test. Generalized one stray "hall-of-helix" example in a
connector.go comment.

FLUID-WP-0008 and FLUID-WP-0009 have no remaining tasks this repo can
act on — the rest are blocked on human steps (BotFather, Telegram
channel creation) or on other repos (fluid-telegram FT-WP-0001 T12,
a second fluid-x/fluid-substack interface) — so both move to status:
blocked.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>

Assistant: claude-code
Assistant-Model: sonnet
Assistant-Process: 251457@bnt-lap001
Assistant-Session: fbcdbf1e-8619-4c11-a73f-2fd270008c35
2026-09-27 22:10:00 +02:00

4.3 KiB

id type title domain repo status flavor owner topic_slug created updated planning_priority planning_order depends_on related_repos state_hub_workstream_id
FLUID-WP-0008 workplan Handover: HelixForge Telegram identity and hall-of-helix channel infotech fluid-core blocked implementation worsch fluid-core 2026-09-04 2026-09-27 high 7
FLUID-WP-0007
fluid-telegram
helix-forge
hall-of-helix
291005a5-d474-5c94-8abd-819f682dde9f

FLUID-WP-0008 - Handover to fluid-telegram

The point of the whole sequence. Establish a Telegram identity for HelixForge and a channel publishing hall-of-helix entries, as the first real FLUID interface. Outbound-first: agents and publishing jobs are the consumers, Telegram is the backend.

Deliverables land in docs/handover/fluid-telegram.md here and as FT-WP-0001 in the fluid-telegram repository.

T01 - Interface evolution intent

id: FLUID-WP-0008-T01
status: done
priority: high
state_hub_task_id: "e3d08f93-836e-56a8-ae20-4acc699e85ef"

Fill the template for helix-forge-telegram-publishing. Parent intent is helix-forge/INTENT.md. Initial authority mode FLUID-2 (advisory): the Daimon may propose and never generate or promote. Out of scope for now: inbound commands, payments, Mini Apps, moderation. Non-authority: never create Telegram entities, change membership, or touch credentials.

T02 - Bot identity — BLOCKED ON A PERSON

id: FLUID-WP-0008-T02
status: todo
priority: high
state_hub_task_id: "9896fee4-c20c-5f49-8428-37c7e221b8cb"

Human step. Register the HelixForge bot via BotFather (Canon BOT-01). Store the token in OpenBao per Canon SEC-04, consistent with HF-WP-0002. The token never enters the repository; the adapter receives only that credential, per Blueprint §28.1 runtime identity.

T03 - Channels — BLOCKED ON A PERSON

id: FLUID-WP-0008-T03
status: todo
priority: high
state_hub_task_id: "bb02bf5b-9d85-5c0d-8b88-9dd67694fb97"

Human step. Create a private test channel first, then the public hall-of-helix channel (Canon PUB-01). Bot is administrator with post rights only. All verification runs against the private channel before anything is published publicly.

T04 - Revision R-1 contract

id: FLUID-WP-0008-T04
status: done
priority: high
state_hub_task_id: "00b2de31-8f84-5478-9489-fedfabf4cd84"

POST /v1/hall-entries (idempotent on entry id) and GET /v1/hall-entries/{id}. OpenAPI plus signed revision descriptor.

T05 - Telegram adapter and connector contract

id: FLUID-WP-0008-T05
status: done
priority: high
state_hub_task_id: "24804694-1581-5d83-a5d4-db46e5af5070"

Map a hall-of-helix/entries/*.md file to a Telegram post. The adapter is a separate process and may be written in whatever suits the platform. Declare the connector capability contract: identity, auth, timeouts, retry, rate limits, error mapping.

T06 - Cohorts, telemetry and redaction

id: FLUID-WP-0008-T06
status: done
priority: high
state_hub_task_id: "231d114a-5f36-53cb-ae81-508566b37d4a"

Subscriber classes and agent consumers as cohorts. Redaction rules that keep Telegram user identity (Canon ID-01) out of the evidence store entirely.

T07 - First competing hypotheses

id: FLUID-WP-0008-T07
status: done
priority: high
state_hub_task_id: "93569266-d282-5c0d-ac72-acbc704d18fd"

Telegram's 4096-character message limit against long-form entries is the first genuine pressure. Two presentation adaptations compete: H-A, teaser plus link back to the hall; H-B, full entry serialized as a multi-part rich post (Canon PUB-02/MED-02). Blueprint §37 argues presentation is the safest and highest-value place to start.

T08 - Live experiment and audit — BLOCKED ON T02, T03 AND SUBSCRIBERS

id: FLUID-WP-0008-T08
status: todo
priority: high
state_hub_task_id: "befd476e-3465-51cd-a850-e006d653f322"

R-2 and R-3 implement H-A and H-B. Run the experiment on real subscriber traffic, measure read-through and reaction, and confirm fluid audit trace reconstructs the whole chain. This is the §50 slice on a real workload.

T09 - Runbook and handover package

id: FLUID-WP-0008-T09
status: done
priority: medium
state_hub_task_id: "be0da67f-67d2-5c0d-bd19-0235310b053e"

Token rotation, channel admin recovery, rollback to R-1, and the assembled handover document.