Skip to content

Commit f97efcf

Browse files
josevalimbenbrandt
authored andcommitted
Add session/load vs session/resume to FAQ
1 parent 3f74e28 commit f97efcf

File tree

1 file changed

+19
-13
lines changed

1 file changed

+19
-13
lines changed

docs/rfds/resume-session.mdx

Lines changed: 19 additions & 13 deletions
Original file line numberDiff line numberDiff line change
@@ -15,25 +15,19 @@ except it does not return previous messages.
1515

1616
> How do things work today and what problems does this cause? Why would we change things?
1717
18-
While the spec provides a "session/load" command, most coding agents do not implement it.
18+
While the spec provides a "session/load" command, not all coding agents implement it.
1919
This means that, once you close your editor, browser, etc, you can't resume the conversation.
2020

21-
Given resuming an existing conversation is already a capability in some agents and genereally
22-
easier to implement in practice, we hope it will be better adopted. Not only that, resuming
21+
This is particularly a problem for agents that do not directly implement ACP and the
22+
functionality is implemented via a wrapper. In such cases, they may provide the ability
23+
to resume (without history), which we would like to hook into. Not only that, resuming
2324
could be used as a mechanism for proxies and adapter libraries to emulate "session/load".
2425

2526
## What we propose to do about it
2627

2728
> What are you proposing to improve the situation?
2829
29-
Two possible options:
30-
31-
* Add a "session/resume" command
32-
* Add an option to "session/load" which allows us to continue even if
33-
it doesn't return the previous conversation
34-
35-
In the rest of the document, I will refer to the solution as "session/resume",
36-
but it implies any of the solutions above.
30+
Add a "session/resume" command and a capability `resumeSession: true`.
3731

3832
## Shiny future
3933

@@ -68,7 +62,7 @@ the stored messages.
6862
### What alternative approaches did you consider, and why did you settle on this one?
6963

7064
The biggest question is if it makes sense to support both "session/load" and
71-
"session/resume". However, I do wonder: how portable "session/load" is in practice?
65+
"session/resume".
7266

7367
When we start a new session over ACP, we introduce custom MCP tools and configuration.
7468
This means that, while we could use "session/load" to load our own chats, loading
@@ -83,6 +77,18 @@ mechanics. If an agent implements "session/load", then it can be used directly,
8377
doesn't, a proxy or adapter can provide a reasonable fallback on top of "session/resume".
8478
This argues "session/resume" is the basic primitive which "session/load" builds on top of.
8579

80+
### Should we introduce a new operation (session/resume) or add an option to (session/load)?
81+
82+
A separate method provides a few benefits:
83+
84+
* for clients that for whatever reason don't want the history, they can just resume
85+
86+
* for agents that can only supply resume, a proxy on top could provide load on top of it,
87+
but the agent is still clear on what it supports
88+
89+
* for agents who support both, it should be trivial to use the same resume functionality,
90+
they just either replay events or do not
91+
8692
## Revision history
8793

88-
<!-- If there have been major updates to this RFD, you can include the git revisions and a summary of the changes. -->
94+
* 2025-11-24: Update FAQ to mention session/resume vs session/load

0 commit comments

Comments
 (0)