poimen-memory/tasks/INDEX.md

poimen-memory — task board

64 tasks — 54 build tasks plus 10 composition gates, one per phase. One file per task, self-contained: inlined design facts, executable steps, acceptance criteria, a Verify section written for someone who did not build the thing, and the traps worth naming. Reading DESIGN.md is not required to do a task — it is linked as background only.

Each Verify section names the harness, the integration test with numbered assertions, the command to run, and the false pass — the shape of test that goes green while the feature is broken. Treat the false-pass list as part of the acceptance criteria, not commentary.

Ordering — declared, never derived

Phase order is the list below. Task ids are opaque and frozen.

M0.3 is M0.3 forever, in whatever phase it currently sits, because its artifacts and cross-references key on that id. New tasks take new ids rather than renumbering neighbours. This is the StepId rule applied to the board itself — a board that renumbers to reorder has the bug it warns its own users about.

No phase starts until its predecessor's gate is green. The gate task at the end of each phase is that gate: it proves the phase's parts compose and that its swappable parts are genuinely swappable. Every build task is verified alone; the gate verifies the properties no single task owns.

Two rules this board exists to protect

1. The JSONL log is authoritative; the vault and the vector index are projections. Anything that cannot be dropped and rebuilt byte-identically from the log has hidden inputs, and that is a bug. M2.8 is the gate that enforces it.

2. The update gate must discriminate, not summarize. Agent transcripts are ~43% tool results and mostly evidence-free. A gate that accepts most chunks is an expensive summarizer that will reproduce the memory-explosion failure the whole design exists to avoid. M1.8 is the gate that enforces it, and update-rate is the single number to watch.

Verification practice — script first, source second

A task is verified by running its command and reading the output, then opening the source. Reviewing the diff first is how an assertion that was quietly dropped still gets called done: the code looks right, and nothing proves the test ran.

Numbered assertion N in a Verify section is test fn aN_<slug>. The numbering is the contract — a test fn that does not exist should report missing rather than be silently absent from a green summary.

cargo test reporting ok with zero tests run is not a pass.

Progress

Source of truth is the Status field in each task file. The tables below mirror it; a status changed here and not there is a lie.

Legend: ⬜ not started · 🟡 in progress · ✅ done · ⛔ blocked

#	Phase	Ids	Tasks	✅	🟡	⬜	Gate
1	Read-only spine	M0.x	8	8	0	0	✅ M0.8
2	Gated loop at L1	M1.x	8	8	0	0	✅ M1.8
3	Projections	M2.x	8	6	0	2	⬜ M2.8
4	L2 synthesis + retrieval	M3.x	4	4	0	0	✅ M3.4
4.5	Distributed API Layer	M3.5.x	10	9	0	1	✅ M3.5.8
5	Skills	M4.x	3	2	0	1	⬜ M4.3
5.5	Reference corpora	M3.6.x	6	1	0	5	⬜ M3.6.6
5.6	Tool context	M3.7.x	6	0	2	4	⬜ M3.7.6
6	Post-training	M5.x	6	0	0	6	⬜ M5.6
7	agent-manager migration	M6.x	6	0	0	6	⬜ M6.6
8	Source connectors	M7.x	10	0	0	10	⬜ M7.10
9	Hybrid search	M8.x	9	1	0	8	⬜ M8.9
	Total		73	49	2	22	5/11 green

Current status — 2025-01-28. Completed phases M0.x, M1.x fully archived (16/16 tasks). M2.1-2 ✅, M2.4-6 ✅ (embeddings, CNPG, pgvector, vault, rebuild). M2.3 ✅ schema, M2.7 ⬜ remains for gate. M3.x (4/4 ✅), M3.5.x (9/10 ✅, M3.5.10 JWT complete). M4.1-2 Skills ✅ done (skill drafting + derived filter). M3.6.1 DocCorpusSource ✅. M8.1 ✅ OpenSearch cluster deployed with Dashboards UI (2-node HA, 30Gi storage, NetworkPolicy, JWT realm TODO for production). Memory Service API upgraded: vault JSON endpoints + hybrid search (semantic 60% + lexical 40%, graceful fallback). All completed task files archived from /tasks/ folder. INDEX.md cleaned to reflect active work only. Significant early work for M3.7: mem-core/src/lesson.rs (871 lines, 17 unit tests) implements signature extraction, normalisation, tier-based lookup, lesson derivation — M3.7.7, M3.7.5 are 🟡. mem-cli/src/lessons_cmd.rs (311 lines), mem-ingest/src/derived_filter.rs (220 lines) provides working mem capture|resolve|lookup|materialize. Tests: 247 passing, 2 ignored (M2.1 +8 tests). Ready to tackle M2.2-8 (projections), M4.3 gate (skills composition), M5 (post-training), M7 (source connectors).

M2.2 (CNPG manifest), M5.4 (vLLM+LoRA), M3.5.9 (git refs), and M6.x (agent-manager migration) are homelab/infra work independent of prior phases, can start parallel.

M6 is a different repo, not a dependency of M0-M5. It migrates github.com/Riotpiaole/agent-manager's session store (a separate Go CLI tool, unrelated to this project's own memory system) from local sqlite to its own dedicated CNPG cluster. It rides in this board because it's homelab work happening alongside M2.2/M5.4, and because the two projects' Postgres schemas landing in the same cluster around the same time need to look like siblings, not strangers — see M6.6's convention-consistency check.

✅ Archived Phases

Completed and archived: M0.x (8/8), M1.x (8/8) — all task files deleted from /tasks/ after verification. See git log for historical record and CLAUDE.md for session context.

3 — Projections · M2.x

Status: In progress · 6/8 done, 2 pending (M2.7, M2.8 gate).

M2.1 ✅ (embeddings client: 768-dim batching @32) M2.2 ✅ (CNPG Cluster + Database CRD with pgvector 0.7.0) M2.3 ✅ (schema + sqlx migrations — M2.3 spec tables, 5 entity types) M2.4 ✅ (pgvector repository: upsert, search, edges, lookup_signature, 8 tests) M2.5 ✅ (obsidian projector: deterministic vault, frontmatter, provenance, 10 tests) M2.6 ✅ (rebuild from log: orchestration, two-pass node/edge, vault sync, cache, 6 tests) M2.7 ⬜ (verify edges — edge closure proof) M2.8 gate awaits M2.7.

4 — L2 synthesis and retrieval · M3.x

Status: ✅ Complete · 4/4 tasks done. Gate M3.4 passing. Task files archived.

4.5 — Distributed API Layer · M3.5.x

Homelab frontend integration: HTTP facade via api.riotpiao.com. Runs in parallel with M4 and M5 after M3.4 green.

Status: 9/10 done · M3.5.8 gate ✅ passing. M3.5.1–8 archived (task files deleted). M3.5.9 (git-aware refs) and M3.5.10 (JWT/OIDC auth) remain. M3.5.10 implementation ✅ complete: Authentik OIDC provider, RS256 validation, capability-based access control. Awaiting new Docker image rollout to pods.

5 — Skills · M4.x

Status: 2/3 done · M4.1 (skill draft) ✅ and M4.2 (derived filter) ✅ archived. Pending M4.3 gate composition.

mem skill draft --project <proj> --from <query-id> writes to vault/skills/_drafts/. Dry-run mode supported. Shingle matcher (M4.2) provides derived LessonSource filtering. M4.1-2 task files deleted after archival.

5.5 — Reference corpora · M3.6.x

Documentation the local models are weak at — kubectl, tea — made retrievable as level R: embedded and indexed, never evidence. Ids are M3.6.x and stay M3.6.x; the phase sits here rather than at 4.6 because M3.6.4 extends M4.2's matcher instead of duplicating it, and because skills are the better answer to the same problem and should exist first.

The load-bearing property is a negative one. Adding a corpus must not change update-rate, must not change default query output, and must not put an R node in any provenance chain. R bypasses the recurrence structurally — run_loop needs a Query and a corpus has none — not by a flag. M3.6.6 asserts M1.8's numbers are unchanged, not merely still-passing, because documentation fed to the gate would lower update-rate and make M1.8 easier to clear while the memory got worse.

Task	Title	Size	Flags	Status
M3.6.1	DocCorpusSource + heading chunking	M	—	✅
M3.6.2	Level R: log, index, vault, rebuild parity	M	—	⬜
M3.6.3	mem ref — replace-on-change corpus management	M	—	⬜
M3.6.4	Reference text cannot re-enter as evidence	M	—	⬜
M3.6.5	Query: filter-then-recall, R opt-in, floor	M	—	⬜
M3.6.6	M3.6 composition gate	M	gate	⬜

5.6 — Tool context · M3.7.x

Answers "what do we already know about this failure, tool or task" over HTTP. Consumers are pi, curl, or an MCP call — nothing here executes a tool, and nothing here serves a tool catalog, because every caller already holds its own MCP schemas.

Three tiers, cheapest first. An exact hash hit on a normalised failure signature means this happened here before; a symptom-vector match means something similar did; the R corpus means nobody here has hit this, read the docs. The tier is a field in the response, because those three answers must not arrive in the same register.

The two tasks that make it work are the least obvious ones. M3.7.7 decides whether tier 1 ever fires — if normalisation leaves a timestamp in, the same failure never hashes twice and the system silently degrades to vector search. M3.7.8 closes the gap between memories written as answers and queries that arrive as stack traces. Both fail invisibly, which is why M3.7.6 ablates them rather than trusting an end-to-end green.

Ids are M3.7.x and frozen. M3.7.1 and M3.7.2 were a tool-catalog surface, deleted before implementation once the consumer was settled; their ids are retired rather than reused.

Task	Title	Size	Flags	Status
M3.7.3	GET /memory/skills?task= — match a subset	M	—	⬜
M3.7.4	/memory/context — three-tier lookup	M	—	⬜
M3.7.5	tool-failures standing query	M	—	🟡
M3.7.6	M3.7 composition gate	M	gate	⬜
M3.7.7	Failure signature extraction + normalisation	M	—	🟡
M3.7.8	Symptom projection at ingest	M	—	⬜

6 — Post-training · M5.x

Python, separate from the Rust workspace. The boundary is the JSONL log.

Task	Title	Size	Flags	Status
M5.1	mem label — evidence labeler	M	—	⬜
M5.2	Labeler calibration	M	—	⬜
M5.3	Training corpus export	M	—	⬜
M5.4	vLLM + --enable-lora	L	homelab	⬜
M5.5	verl training loop	L	—	⬜
M5.6	M5 composition gate	L	gate	⬜

7 — agent-manager migration · M6.x

Separate repo (github.com/Riotpiaole/agent-manager, fork branch add-headless-spawn), separate cluster resource, no Rust/GRU-Mem dependency. Moves its session store off local sqlite onto a dedicated CNPG Postgres, reachable from the Mac client through a dedicated nginx route — durability-of-location, not a multi-host requirement.

Task	Title	Size	Flags	Status
M6.1	CNPG agent-manager-db manifest	M	homelab	⬜
M6.2	Postgres schema for agent-manager sessions	M	—	⬜
M6.3	store.go query port to Postgres	L	—	⬜
M6.4	nginx TCP routing to agent-manager-db	S	homelab	⬜
M6.5	Postgres credentials for the Mac client	S	homelab	⬜
M6.6	M6 composition gate	M	gate	⬜

8 — Source connectors · M7.x

Extensible multi-source ingestion. SourceConnector trait + YAML-driven registry. Adding a new document source (paperless-ngx, S3, git repo) requires implementing one trait and adding one config block — no changes to the ingest pipeline, chunking, embedding, storage, or query layers.

Document connectors produce Level R content (reference material, bypasses gated loop). Session connectors (pi, claude) produce evidence for L0/L1/L2. The connector's source_type() declares the pipeline.

Sync framework handles change detection (sha-based skip), tombstoning, drift reporting, and resumable sync for all connectors.

Task	Title	Size	Flags	Status
M7.1	SourceConnector trait + registry	M	—	⬜
M7.2	Obsidian vault connector	M	—	⬜
M7.3	paperless-ngx connector	M	—	⬜
M7.4	Git repo connector	M	—	⬜
M7.5	S3-compatible storage connector	M	—	⬜
M7.6	Sync framework	L	—	⬜
M7.7	mem source CLI	M	—	⬜
M7.8	Source HTTP endpoints	M	—	⬜
M7.9	Connector health + observability	S	—	⬜
M7.10	M7 composition gate	M	gate	⬜

---

9 — Hybrid search · M8.x

Parallel retrieval from pgvector (semantic) and OpenSearch (lexical), fused with Reciprocal Rank Fusion. Adds a QueryOptimizer that classifies queries and routes to the best strategy before any database call.

The load-bearing property is accuracy. Hybrid must produce measurably better NDCG@10 than either engine alone. If it doesn't, the gate fails — not because the code is broken, but because the system isn't earning its complexity budget. M8.9 requires benchmark numbers, not just green tests.

Approach A: Parallel RRF. Both engines run simultaneously via tokio::try_join!. Results are merged by rank position, not score magnitude, because pgvector cosine ([0,1]) and BM25 ([0,50+]) are incomparable distributions. RRF needs no parameter tuning (k=60 is the academic standard). The alternative — weighted linear combination — requires labelled data for weight selection that we don't have yet.

Dual-write indexing. Every chunk gets the same UUID in both pgvector and OpenSearch. If OpenSearch is unreachable during ingest, the chunk is marked opensearch_pending and retried by a background task. The gate (M8.9) checks for zero orphans.

Fallback. If OpenSearch is down at query time, the worker degrades to semantic-only. If the embedding model is down, it degrades to lexical-only. The response search_strategy field always reports which mode was actually used.

Task	Title	Size	Flags	Status
M8.1	OpenSearch cluster + JWT realm	M	homelab	✅
M8.2	Dual-write indexing pipeline	M	—	⬜
M8.3	Query optimizer: context + routing	M	—	⬜
M8.4	Reciprocal Rank Fusion engine	S	—	⬜
M8.5	Hybrid query worker: parallel retrieval	L	—	⬜
M8.6	Upgrade GET /query to hybrid + fallback	M	—	⬜
M8.7	Index tuning: HNSW + OpenSearch analyzers	M	—	⬜
M8.8	Accuracy benchmarks: NDCG, MRR, P@K	M	—	⬜
M8.9	M8 composition gate	M	gate	⬜

---

Background: DESIGN.md · GRU-Mem, arXiv 2602.10560 · internal/store/store.go (agent-manager, add-headless-spawn branch)