Skip to content

docs(adr): synchronous workflow_call for event-driven agent dispatch#1245

Open
ifireball wants to merge 1 commit into
fullsend-ai:mainfrom
ifireball:docs/adr-0039-synchronous-workflow-call
Open

docs(adr): synchronous workflow_call for event-driven agent dispatch#1245
ifireball wants to merge 1 commit into
fullsend-ai:mainfrom
ifireball:docs/adr-0039-synchronous-workflow-call

Conversation

@ifireball
Copy link
Copy Markdown
Contributor

Summary

  • Propose ADR 0039: replace dispatch.yml → stage workflow_dispatch / gh workflow run with synchronous workflow_call so Actions runs stay linked for PR/issue debugging.
  • Restore org custom-agent discoverability via # fullsend-stage: metadata, a scaffolded sync-dispatch script, skills, and CI in .fullsend (follow-on implementation).

Test plan

  • make lint on ADR 0039
  • ADR review only — no workflow or runtime changes in this PR

Made with Cursor

Record ADR 0039 to replace dispatch.yml gh workflow run fan-out with
workflow_call for UI traceability, and compile-time dispatch sync for
org custom agents in .fullsend.

Co-authored-by: Cursor <[email protected]>
@github-actions
Copy link
Copy Markdown

Site preview

Preview: https://8a28cfed-site.fullsend-ai.workers.dev

Commit: ae372965e71baa66f132ca2fdeb67d39fa5e31b1

@ifireball ifireball self-assigned this May 20, 2026
@fullsend-ai-review
Copy link
Copy Markdown

fullsend-ai-review Bot commented May 20, 2026

Review

Findings

No findings.

Previous run

Review

Findings

No findings.

@fullsend-ai-review fullsend-ai-review Bot added the ready-for-merge All reviewers approved — ready to merge label May 20, 2026
@ifireball ifireball marked this pull request as ready for review May 20, 2026 10:26
@fullsend-ai-review fullsend-ai-review Bot added ready-for-merge All reviewers approved — ready to merge and removed ready-for-merge All reviewers approved — ready to merge labels May 20, 2026
Copy link
Copy Markdown
Contributor

@rh-hemartin rh-hemartin left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I agree.

Did we take into account the billing for GH actions? There limits are very large in terms of minutes, but I don't know if this would be doubling minute usages for orgs using fullsend.

As a side note it is always funny how a lot of ADR frame options are very one sided.

Copy link
Copy Markdown
Contributor

@ggallen ggallen left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This reverses the decision made in ADR-0026, taking away what I feel is essential functionality.

If this is done then replacing or adding new agents becomes much more complicated.

@ifireball
Copy link
Copy Markdown
Contributor Author

ifireball commented May 25, 2026

@rh-hemartin

I agree.

Did we take into account the billing for GH actions? There limits are very large in terms of minutes, but I don't know if this would be doubling minute usages for orgs using fullsend.

As a side note it is always funny how a lot of ADR frame options are very one sided.

Not quite sure why are you claiming this particular ADR would contribute towards more usage minutes, it could end up being cheaper if e.g. workflow_call is cheaper then workflow_dispatch.

@ifireball
Copy link
Copy Markdown
Contributor Author

@ggallen

This reverses the decision made in ADR-0026, taking away what I feel is essential functionality.

If this is done then replacing or adding new agents becomes much more complicated.

Yes, I'm being explicit about that, trying to find a middle ground where we would add build time tooling and CI to do "agent discovery" instead a fully dynamic discovery which requires async dispatch.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

ready-for-merge All reviewers approved — ready to merge

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants