[feat] Stop evicting warm sessions when an integration is added - #6368
Merged
Conversation
|
The latest updates on your projects. Learn more about Vercel for GitHub.
|
dosubot
Bot
added
size:L
Contributor
📘 Docs preview
This comment updates in place on every push. |
Contributor
|
Important Review skippedAuto reviews are disabled on base/target branches other than the default branch. 🗂️ Base branches to auto review (1)
Please check the settings in the CodeRabbit UI or the ⚙️ Run configurationConfiguration used: Repository YAML (base), Organization UI (inherited) Review profile: CHILL Plan: Pro Plus Run ID: You can disable this status message by setting the Use the checkbox below for a quick retry:
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
mmabrouk
added
lgtm
Adding a second integration to an agent evicted its warm session twice over: the integration names were composed into the prompt strings (appendSystemPrompt for Pi, agentsMd for the file-based harnesses) and into the search_tools tool description, and both surfaces are part of the session fingerprint. A one-word list change cost a full sandbox rebuild, fresh Daytona Secrets, and the substitution race of #6362. Now the guidance rides the wire as gatewayGuidance {text, carrier}: - the SDK adapters keep the carrier choice but stop splicing; the prompt strings leave the adapter purely authored; - the runner splices guidance-first at environment build (buildRunPlan), so the text refreshes exactly when a session is built and never evicts one: configFingerprint and the desired-state facets exclude the field by design, with the reason documented at both sites; - the wording presents the names as examples ('For instance, some of the integrations you have: ... Others may exist, and this list can go stale — search_tools is the source of truth'), so a list that goes stale mid-session stays honest; - the derived search_tools/run_tool descriptions drop the names sentence and are now byte-identical for any integration set, so only the FIRST connection (which genuinely adds the two tools) changes session config. Golden run_request.gateway_connection.json pins the new field; the connection-free payloads are byte-identical to before. Claude-Session: https://claude.ai/code/session_014s6jqKCsVKsNMmnrJVJkZt
mmabrouk
force-pushed
the
chore/remove-pi-agenta
branch
from
August 31, 2026 15:44
fa9b103 to
2c96b8e
Compare
mmabrouk
force-pushed
the
feat/gateway-guidance-field
branch
from
August 31, 2026 15:44
b1503f1 to
3c21ca2
Compare
mmabrouk
changed the base branch from
chore/remove-pi-agenta
to
release/v0.114.4
August 31, 2026 16:36
Merged
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Context
Adding a second integration to an agent evicted its warm session, twice over. The integration names were written into the prompt guidance ("Configured integrations: github, slack.") and into the
search_toolstool description, and both surfaces are hashed into the session fingerprint. So a one-word list change cost a full sandbox rebuild, fresh Daytona Secrets, and a pass through the substitution race of #6362. Editing an integration's tool permissions was already fine:gatewayPolicysits outside the fingerprint.Changes
The guidance becomes its own wire field, and the names become examples.
Before: the SDK adapter spliced the guidance into
appendSystemPrompt(Pi) oragentsMd(Claude, Codex), and the runner hashed the result.After: the wire carries
gatewayGuidance: {text, carrier}. The adapters keep the carrier choice but stop splicing, so the prompt strings leave the SDK purely authored. The runner splices guidance-first when it builds an environment (buildRunPlan), andconfigFingerprintplus the desired-state facets exclude the field by design, with the reason documented at both sites. The text refreshes exactly when a session is built, and never evicts one.The wording changes with it: "For instance, some of the integrations you have: github, slack. Others may exist, and this list can go stale —
search_toolsis the source of truth." A list that goes stale mid-session is now honest.The derived
search_tools/run_tooldescriptions drop their names sentence and are byte-identical for any integration set. So only the first connection, which genuinely adds the two tools, changes session config at all.Tests
run_request.gateway_connection.jsonpins the new field. Connection-free payloads are byte-identical to before.gatewayGuidancehash equal.What to QA
search_toolsimmediately; its prompt still names only the first until the next cold start.Stacked on #6367.
https://claude.ai/code/session_014s6jqKCsVKsNMmnrJVJkZt