Brokered comments should contain all the logs emitted for the brokered action that created the comment #1079

New issue

Open

opened 2026-07-10 20:42:04 +00:00 by coilysiren · 0 comments

coilysiren commented

2026-07-10 20:42:04 +00:00

Owner

The problem

This issue asks for similar thing #1078

An important component here - preflight itself a brokered action (perhaps the most important one) so it should set the pattern here. Right now the preflight reservation comment just includes some metadata. This expands the read more section of that comment to include all the logs. Except the verbatim issue text, as that would be recursive. And by all the logs I truly mean all of them, like the substrate hydration logs, the docker pull logs, etc. a new method of logs capture might need to be a setup to meet this goal.

This should capture as much as possible. The natural thing you will reach for is to post the preflight logs from the host exec. That's not what I want. The goal is to capture everything up until immediately prior to LLM handoff. The best way to do this might be to post in 2 parts - start the comment on the host (make sure to indicate that it's pre flight stage) then finish the comment at the end of container internal bootstrap. Or maybe 2 comments. Probably 2 comments.

The finalization of an issue / run should got a similar treatment

Proposed change

Alternatives considered

Before filing

I searched existing issues and this is not a duplicate.
This stays within ward's scope (the dev-verb gate / agent driver), not a personal-infra or downstream-repo verb.

### The problem This issue asks for similar thing https://forgejo.coilysiren.me/coilyco-flight-deck/ward/issues/1078 An important component here - preflight itself a brokered action (perhaps the most important one) so it should set the pattern here. Right now the preflight reservation comment just includes some metadata. This expands the read more section of that comment to include *all* the logs. Except the verbatim issue text, as that would be recursive. And by all the logs I truly mean all of them, like the substrate hydration logs, the docker pull logs, etc. a new method of logs capture might need to be a setup to meet this goal. This should capture as much as possible. The natural thing you will reach for is to post the preflight logs from the host exec. That's not what I want. The goal is to capture everything up until immediately prior to LLM handoff. The best way to do this might be to post in 2 parts - start the comment on the host (make sure to indicate that it's pre flight stage) then finish the comment at the end of container internal bootstrap. Or maybe 2 comments. Probably 2 comments. The finalization of an issue / run should got a similar treatment ### Proposed change // ### Alternatives considered // ### Before filing - [x] I searched existing issues and this is not a duplicate. - [x] This stays within ward's scope (the dev-verb gate / agent driver), not a personal-infra or downstream-repo verb.