fix(coding-agents): stop promising automatic correction supersession - #4205
Closed
PaleRoses wants to merge 1 commit into
Closed
fix(coding-agents): stop promising automatic correction supersession#4205PaleRoses wants to merge 1 commit into
PaleRoses wants to merge 1 commit into
Conversation
Collaborator
|
yes it does because the consolidation will update the existing Observations, that's why it's all based on Knowledge Pages which defaults to Observations only! |
Author
|
You're right. This PR generalizes ingestion-time behavior to the full correction lifecycle and misses the observation-update/page-refresh path. |
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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
The coding-agent SessionStart preamble says that newer facts supersede older ones. Its skill promises that one correction permanently outranks the stale memory. Ingesting a correction does not establish either guarantee: it adds evidence, but does not directly rewrite the stored knowledge page or guarantee retrieval precedence. An agent can therefore leave the current answer wrong while believing that ingest alone repaired it.
Change
Three files only: skill source, generated skill, and src/core/knowledge-injection.ts. No storage, retrieval, consolidation, API, stale-page flags, or ingestion-wait behavior changes. Independent of #4170.
Verification
On an isolated worktree based on upstream 9e6d9e7:
No live LLM calls or backend behavior claims. Prepared with automated assistance; the commands above were executed against this PR's tree.