docs(eda): resolve the Channel action position

Notification is a contract in internal/domain/messaging, not a channel type.
A channel implementation satisfies it by importing its own parent, which it
already does, so channels.Channel gains no notification method and no import
and the dependency stays channels -> messaging. A notification has no
lifecycle, no inbound direction and no config schema, so modelling it as a
channel would force Start/Stop/ConfigSchema onto something with no meaning
for them.

Destinations are configured, never named by a playbook: [notify] grows from
one webhook URL into named destinations, each binding a name to a channel
already registered in the Messaging composition plus that channel's address.
An action names a destination. Playbook triggers are reachable from a public
intake surface, so an action that could address an arbitrary chat would be a
new exfiltration surface.

Host access needs no new process surface. archie-messaging dials out only and
deliberately has no NATS connection, so nothing can push to it today.
Delivery travels the Gateway connection every process already holds, as a new
server-streaming RPC on ChatService; the per-turn Stream RPC is not reused.

The 2026-09-03 section presenting two options is superseded. Its claim that no
send capability exists anywhere was too strong: telegram.SendPendingReport
already sends unsolicited to a stored chat id, with no interface for it.

Closes archie-core-t2db.19
This commit is contained in:
sam committed 2026-09-22 00:06:52 +10:00
1 parent 1f6ba21d9a
commit 44dde18918
1 file changed
+97 -57
+97 -57
View File
@@ -1,13 +1,13 @@
# Channel and Forge action positions -- investigation (resolves `t2db.19`)
# Channel and Forge action positions (resolves `t2db.19`)
**Status:** Investigation complete for Forge; Channel corrected and flagged
for a decision only Sam can make -- not a design recommendation.
**Date:** 2026-09-03
**Status:** Draft
**Date:** 2026-09-03, Channel resolved 2026-09-22
**Parent:** `docs/prds/eda-playbook-engine.md`, epic `archie-core-t2db`
**Blocked on:** `archie-core-t2db.17` (idempotency keying) for implementation
of either position -- both are side-effecting.
**Blocked on:** `archie-core-t2db.24` (the `playbook_dispatches` ledger) for
implementation of either position. Both are side-effecting, and `t2db.17`
resolved the keying scheme in design only.
## Forge: ready to design once t2db.17 lands
## Forge: ready to implement once the ledger lands
`internal/forge.Forge` is already exactly the shape a native-Go action
position needs -- a real, typed, general-purpose interface, not something
@@ -50,8 +50,8 @@ layer today. Depend on `Forger`, not on `internal/forge`.
starting set. `create-pr` is explicitly a heavier case -- opening a PR is a
harder-to-reverse action with its own dedup shape beyond "did this event
fire before" (a redelivered event must not open a second PR for the same
logical change) -- defer it to a follow-up slice once `t2db.17`'s keying
scheme is proven on the simpler operations, don't design it blind now.
logical change) -- defer it to a follow-up slice once the keying scheme is
proven on the simpler operations, don't design it blind now.
**Host access:** the coordinator needs the same `Forge` instance
`internal/daemon`'s poller already holds (constructed once at startup via
@@ -59,65 +59,97 @@ scheme is proven on the simpler operations, don't design it blind now.
existing instance into the coordinator's constructor), not a new
extensibility mechanism.
## Channel: the crew's assumption doesn't hold -- flagged, not designed
## Channel: resolved 2026-09-22
The investigation backlog that recommended scoping this alongside Forge
assumed `channels.Channel` already has an addressable send capability that
just needs a name-based lookup. **That's not true.** Verified directly:
Superseded: the 2026-09-03 section here presented two options and asked for a
decision. The decision is below; the options are not repeated.
### Constraints this satisfies
Messaging is the root, Channel is the integration layer, an implementation
such as Telegram sits under Channel. A channel is added by implementing one
reusable interface, and every channel must implement it. A notification
mechanism belongs under Messaging, and `channels.Channel` must not reach
across into it.
### Notification is a Messaging contract, not a channel type
The contract lives in `internal/domain/messaging`. A channel implementation
satisfies it by importing its own parent, which it already does, so the
dependency stays `channels -> messaging` and nothing points back.
`channels.Channel` does not change. It gains no notification method and no
import, so the interface every channel must implement stays the lifecycle and
config contract it is today. `internal/channels/telegram` satisfying both
contracts is an implementation choosing to, not the Channel layer knowing
about notification.
A notification is not a channel type because it has no lifecycle to start and
stop, no inbound direction, and no configuration schema of its own. Modelling
it as one would force `Start`/`Stop`/`ConfigSchema` onto something with no
meaning for them, and make every real channel declare whether it is a
notification.
### Destinations are configured, never named by a playbook
`[notify]` today is one webhook URL, read in one place
(`internal/domain/workflow/feasibility.go:134`). It becomes a set of named
destinations. Each binds a name the operator chooses to a channel already
registered in the Messaging composition (`internal/app/archiemessaging/compose.go`
registers `telegram`, `email`, `webhook`) plus that channel's own address,
such as a Telegram chat id. The existing single-webhook form keeps working as
a destination named `webhook`.
A playbook action names a destination. It never names an address. Playbook
triggers are reachable from a public, unauthenticated intake surface, so an
action that could address an arbitrary chat would be a new exfiltration
surface; a destination the operator configured once is not.
### Args and Result
```go
// internal/gateway/gateway.go
type Gateway interface {
Name() string
Start(ctx context.Context, router *Router, lifecycle Lifecycle) error
Stop(ctx context.Context) error
type NotifyArgs struct {
Destination string // a name from [notify]
Text string
}
type NotifyResult struct {
Delivered bool
}
```
`channels.Channel` embeds exactly this plus `ConfigSchema`/`ValidateConfig`
-- lifecycle and config introspection, nothing about sending a message.
Grepping `internal/channels` and `internal/gateway` for any `Send` method
returns nothing. The only generic outbound-message mechanism in the entire
codebase is `Notify` (`internal/config`'s `[notify]` block) -- a single
fixed webhook URL for gate failures/approvals/parked tasks, not a
per-channel-addressable send.
An unknown destination is a reported error at playbook load, not at dispatch:
the destination set is known at startup, so this follows the engine's
reject-at-load rule.
Real per-channel sending exists, but deliberately not as a reusable
capability: `telegramApprover` (`internal/channels/telegram/approval.go`)
captures a `*bot.Bot` plus a specific `chatID`/`threadID`/`recipient` at
construction, scoped to one conversation for one launch. This is the
established precedent in this codebase (already in project memory from
before this session) precisely to prevent reading a shared bot handle
across chats -- a playbook-triggered send has no chat turn to be scoped to,
so this precedent doesn't transfer as-is.
### Host access: no new process surface
**This means "Channel action position" is not a small wrapper over an
existing capability -- it requires deciding whether to build a genuinely
new send capability that doesn't exist today, and if so, what it targets.**
Two different things could reasonably be meant by "Channel" here, and they
have different shapes:
`archie-messaging` is its own process (`cmd/archie-messaging`). It dials out
to the Gateway and the State Store and serves nothing, and it deliberately
has no NATS connection: `internal/app/archiemessaging/telegram_features.go:65`
records that it "must not decode" the daemon's `[nats]` config. So a
notification cannot be pushed into it today by any route.
1. **A new, playbook-scoped notification capability** -- closer in spirit
to `Notify` (fire a message somewhere, no chat-turn context needed) but
addressable by name/target instead of one fixed webhook. This would be
new engine-family work (its own typed contract, registry, narrow host
access), not an extension of `channels.Channel`.
2. **Reusing the interactive chat channels** (Telegram, email) for
playbook-originated sends -- requires deciding what conversation context
a playbook-triggered message has (which chat? whose approval flow?) when
there was no incoming chat turn to anchor it, and building whatever
capture/scoping mechanism the interactive channels use today, generalized
to a non-chat trigger. This is materially harder and touches the
gateway's session model.
Delivery therefore travels the connection that already exists. Every process
that needs Messaging already dials the Gateway: `archied`
(`internal/app/archied/chat_service.go:25`), `archie-messaging`
(`internal/app/archiemessaging/run.go:26`) and `archie-ui`. The Gateway is the
hub. A new server-streaming RPC on `ChatService` lets `archie-messaging`
subscribe at startup and receive notifications the daemon publishes; Messaging
then delivers through the named channel instance it already holds in
`Service.channels`. No new listener, no NATS in Messaging, no change to how
any process is reached.
**This is not a call to make unilaterally.** It changes what "Channel
action" even means, and the answer affects the trust/scoping model, not
just an implementation detail. Recommend: Sam decides which of the two (or
a third option not listed here) before any Channel position design work
starts. Until then, `t2db.19` should be treated as **Forge scoping only**
-- the Channel half stays genuinely open, not just unimplemented.
The existing `Stream` RPC is not reused: it streams one chat turn's events for
one inbound message (`internal/infrastructure/gatewayrpc/server.go:Stream`).
## Packages this touches (Forge only, once `t2db.17` lands)
### Out of scope
Giving playbook-originated sends the session and approval scoping the
interactive channels use. Configured destinations carry the address, so there
is no conversation to anchor to and no scoping mechanism to generalize.
## Packages this touches
- `internal/domain/eda` (or a new sibling package, e.g.
`internal/domain/eda/forgeaction`): typed `Args`/`Result` per operation,
@@ -128,3 +160,11 @@ starts. Until then, `t2db.19` should be treated as **Forge scoping only**
the daemon's already-built `forge.Forge` client, passed in as the
`workflow.Forger` it already satisfies -- no new lifecycle to manage,
and no change to `internal/forge` itself.
- `internal/domain/messaging`: the notification contract and its
`NotifyArgs`/`NotifyResult`.
- `internal/config`: `[notify]` becomes named destinations.
- `proto/gateway/v1` and `internal/infrastructure/gatewayrpc`: the
server-streaming notification RPC.
- `internal/app/archiemessaging`: subscribe at startup, resolve a destination
to a registered channel instance, deliver.