1e97fb9a16
Self-review (post-merge) flagged that the backfill claimed to be
idempotent on re-run but actually duplicates every row because the
plugin's INSERT uses gen_random_uuid() and ignores any id passed in.
Fix is contract-level: extend MemoryWrite with an optional `id`
idempotency key. When supplied, the plugin MUST treat the write as
upsert keyed on this id; when omitted, the plugin generates a fresh
UUID (production agent commits keep working unchanged).
Changes:
* docs/api-protocol/memory-plugin-v1.yaml: add id field with
description that flags it as idempotency key
* internal/memory/contract/contract.go: add ID to MemoryWrite struct,
update memory_write_minimal golden vector
* internal/memory/pgplugin/store.go: split CommitMemory into two
paths — upsert when body.ID set (INSERT ... ON CONFLICT (id) DO
UPDATE), plain INSERT otherwise
* cmd/memory-backfill/main.go: pass agent_memories.id to MemoryWrite,
fix the false comment about 409 deduplication
New tests:
* pgplugin: TestCommitMemory_WithIDUpserts pins the upsert SQL is
used when id is set; TestCommitMemory_UpsertScanError covers the
error branch
* backfill: TestBackfill_PassesSourceUUIDAsIdempotencyKey pins the
forwarding behavior; TestBackfill_RerunIsIdempotent simulates a
retry and asserts both runs pass the same uuid (plugin upsert is
what makes this safe)
Why this matters: operators retrying a failed backfill (which they
will — networks fail, transactions abort) would otherwise create N
duplicates per memory. The duplicates aren't visible until search
results show obvious dupes — debugging that under prod load is bad.
Production agent commits are unaffected: they leave id empty, the
plugin generates a fresh UUID via gen_random_uuid(), zero behavior
change for the hot path.
|
||
|---|---|---|
| .. | ||
| adapters | ||
| adr | ||
| agent-runtime | ||
| api-protocol | ||
| architecture | ||
| assets | ||
| blog | ||
| development | ||
| devrel/demos/tool-trace-platform-instructions | ||
| engineering | ||
| frontend | ||
| guides | ||
| incidents | ||
| infra | ||
| integrations | ||
| memory-plugins | ||
| pages/api | ||
| plugins | ||
| tutorials | ||
| .gitignore | ||
| api-reference.md | ||
| ecosystem-watch.md | ||
| glossary.md | ||
| index.md | ||
| internal-content-policy.md | ||
| quickstart.md | ||
| README.md | ||
| workspace-runtime-package.md | ||
docs/
This directory serves two purposes:
- Markdown content — everything under
architecture/,agent-runtime/,api-protocol/,development/,frontend/,plugins/,product/, etc. This is what agents and humans read. - VitePress site —
.vitepress/config.ts,package.json,package-lock.json. These drive the rendered documentation site.
Local preview
cd docs
npm install
npm run dev # preview on http://localhost:5173
npm run build # static build to docs/.vitepress/dist/
Conventions
- New top-level docs must be linked from
PLAN.md,README.md, andCLAUDE.md— otherwise agents can't find them (see.claude/memoryfeedback_cross_reference_docs.md). edit-history/YYYY-MM-DD.mdis append-only log of significant changes; don't rewrite history.archive/holds one-shot analyses and retired docs — kept for context but not maintained.
Why site tooling lives here (not in docs-site/)
VitePress expects its config at <root>/.vitepress/config.ts where <root> is also the content directory. Splitting tooling into a sibling docs-site/ would require a non-trivial srcDir shim and break relative links in .vitepress/config.ts. Keeping both together is the pragmatic choice; this README is the tradeoff ledger.