feat(ai): expose prompt config on fetched prompts - #4329
Merged
Conversation
Contributor
|
Reviews (1): Last reviewed commit: "feat(ai): expose prompt config on fetche..." | Re-trigger Greptile |
Contributor
|
Contributor
Contributor
|
Size Change: +1.5 kB (+0.01%) Total Size: 17.9 MB 📦 View Changed
ℹ️ View Unchanged
|
Member
|
the cache-pollution test mutates only top-level properties. Since configs commonly contain |
marandaneto
approved these changes
Jul 30, 2026
marandaneto
left a comment
marandaneto
left a comment
Member
There was a problem hiding this comment.
left a test suggestion but not a blocker, LGTM
This was referenced Jul 30, 2026
Merged
5 tasks
Contributor
Author
|
Fixed in cf01aff. The test now also mutates a nested value, so swapping structuredClone for a shallow copy would fail the test |
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
2 participants
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Problem
PostHog prompt versions can now store a
configobject next to the prompt content (PostHog/posthog#74264): model parameters, tools, or any agent configuration, versioned with the prompt and returned by the fetch endpoint. The SDK parses API responses into typed result objects, so the field is currently fetched and thrown away. Python counterpart: PostHog/posthog-python#801.Changes
PromptRemoteResultgainsconfig: Record<string, unknown> | null, always present on api/cache/stale_cache results (null when the version has none). The fallback result declaresconfig: undefined, matching how it handlesnameandversion.configor send a non-object shape read asnull(same tolerance as the existing "server ignored the label param" handling)structuredClone, so mutatingresult.configbefore spreading it into an LLM call cannot pollute later cache hits (this exact bug was caught by review on the Python PR; here it is prevented from the start)Release info Sub-libraries affected
Libraries affected
Checklist
If releasing new changes
pnpm changesetto generate a changeset file🤖 Agent context
Autonomy: Human-driven (agent-assisted)
Claude Code authored this under my direction as the JS counterpart of PostHog/posthog-python#801. Four new tests (config through api and cache, stale-cache retention, cache-pollution guard, older-server matrix); the four existing full-equality metadata assertions now include
config: null, pinning the "config is always present on remote results" contract. The changeset file was written manually in.changeset/.