feat(memory): Add storage-backed plugin package#623
Draft
dcramer wants to merge 4 commits into
Draft
Conversation
Add the initial memory plugin package with SQL schema, migration artifacts, deterministic write policy, and a scoped store for create, list, search, and archive operations. Wire the package into release automation and add component coverage over the real plugin DB migration path so the storage boundary is reviewable before tools, prompt recall, extraction, or embeddings are introduced. Co-Authored-By: GPT-5 Codex <codex@openai.com>
|
The latest updates on your projects. Learn more about Vercel for GitHub.
|
![cursor[bot]](https://avatars.githubusercontent.com/in/1210556?s=60&v=4){kind=small}
Add explicit memory source kind and supersession timestamp fields so stored rows preserve the provenance and lifecycle data needed by later extraction and supersession work. Keep active duplicate and visibility filters aligned with superseded state. Co-Authored-By: GPT-5 Codex <gpt-5-codex@openai.com>
There was a problem hiding this comment.
Cursor Bugbot has reviewed your changes and found 1 potential issue.
There are 2 total unresolved issues (including 1 from previous review).
❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.
Reviewed by Cursor Bugbot for commit 5f9cfc2. Configure here.
Keep the authoritative memory row focused on concrete storage, visibility, source, idempotency, and lifecycle fields. Remove speculative provenance, metadata, subject label, and extraction confidence columns until the extraction or graph paths introduce concrete consumers. Co-Authored-By: GPT-5 Codex <gpt-5-codex@openai.com>
Align the memory store with the lean storage contract after review. Keep Slack conversation scope source-derived, archive expired duplicates before recreation, exclude expired rows from archive, remove speculative archived-list and policy exports, and update the memory specs to match the first storage slice. Co-Authored-By: GPT-5 Codex <gpt-5-codex@openai.com>
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.
Adds the first memory plugin package slice: a trusted @sentry/junior-memory registration, plugin-owned SQL migration/schema, deterministic write policy, and a scoped store for create, list, search, and archive operations. The store derives personal and conversation visibility from runtime context, supports scoped idempotency, and keeps lexical search inside the visible active predicate.
I checked both the Ash memory prior art and broader memory systems before settling the storage shape. Ash uses an authoritative memories table plus a separate rebuildable memory_embeddings vector index. LangGraph frames long-term memory as scoped documents keyed by namespace/key, Letta keeps editable memory blocks and persisted messages separate, Mem0 exposes id/content/user scope/category/timestamp style memory records, and Zep puts richer entity/fact invalidation into a graph layer. For Junior, that points to a lean source-of-truth row now, with vector storage, graph/entity data, extraction confidence, subject labels, and operational metadata deferred until concrete consumers exist.
The memory row now carries only the fields needed for the current explicit tool/storage slice and near-term lifecycle semantics: authority-bearing scope/source, type/sensitivity, content/hash, idempotency, observed/created/expiration timestamps, and archive/supersession markers. This intentionally stops before model-visible tools, automatic prompt recall, passive extraction, and embeddings. The PR wires the package into release automation and adds component coverage over the real plugin DB migration path, including visibility isolation, policy rejection, idempotent retries, and lexical search beyond newest-row limits.