@plur-ai/mcp 0.14.0 → 0.16.0
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/README.md +12 -16
- package/dist/{server-LAXRBVKA.js → chunk-TPL7QPPO.js} +296 -540
- package/dist/index.js +6 -6
- package/dist/server-VHSO4XMV.js +378 -0
- package/dist/tools-export.d.ts +76 -0
- package/dist/tools-export.js +21 -0
- package/package.json +9 -3
- package/packs/effective-memory/SKILL.md +1 -1
- package/packs/effective-memory/engrams.yaml +3 -3
package/README.md
CHANGED
|
@@ -46,26 +46,24 @@ Knowledge is stored as **engrams** — small assertions that strengthen with use
|
|
|
46
46
|
|
|
47
47
|
## Tools
|
|
48
48
|
|
|
49
|
-
|
|
49
|
+
By default (lean profile), your agent gets 12 tools. Everything else is reachable through `plur_admin`:
|
|
50
50
|
|
|
51
51
|
| Tool | What it does |
|
|
52
52
|
|------|-------------|
|
|
53
53
|
| `plur_session_start` | Start a session — injects relevant engrams for your task |
|
|
54
54
|
| `plur_learn` | Store a memory — correction, preference, convention, or decision |
|
|
55
|
-
| `
|
|
56
|
-
| `plur_recall_hybrid` | **Best default** — BM25 + embeddings merged via RRF. Zero cost. |
|
|
57
|
-
| `plur_recall` | Keyword search (BM25 only, instant) |
|
|
58
|
-
| `plur_inject_hybrid` | Load relevant memories for the current task |
|
|
55
|
+
| `plur_recall` | **Best default** — hybrid (BM25 + embeddings via RRF) by default; pass `mode:"keyword"` for BM25-only. Zero cost. |
|
|
59
56
|
| `plur_feedback` | Rate a memory — trains relevance over time |
|
|
60
57
|
| `plur_forget` | Retire a memory (history preserved) |
|
|
61
58
|
| `plur_session_end` | End a session — captures summary and new learnings |
|
|
62
|
-
| `plur_capture` | Record a session event |
|
|
63
|
-
| `plur_timeline` | Query session history |
|
|
64
|
-
| `plur_ingest` | Extract learnings from text |
|
|
65
|
-
| `plur_sync` | Sync memory across machines via git |
|
|
66
|
-
| `plur_packs_install` | Install a shareable memory pack |
|
|
67
|
-
| `plur_packs_list` | List installed packs |
|
|
68
59
|
| `plur_status` | System health |
|
|
60
|
+
| `plur_doctor` | Diagnose embedder, hybrid search, and remote-store auth |
|
|
61
|
+
| `plur_receipt` | Show why a memory was injected — the evidence behind a recall |
|
|
62
|
+
| `plur_packs_uninstall` | Remove an installed pack |
|
|
63
|
+
| `plur_tensions_purge` | Clear stale/resolved tensions |
|
|
64
|
+
| `plur_admin` | Dispatch to any other tool: `{ action: "plur_packs_install", args: {...} }` |
|
|
65
|
+
|
|
66
|
+
Less commonly needed tools (`plur_recall_hybrid`, `plur_inject_hybrid`, `plur_learn_batch`, `plur_ingest`, `plur_sync`, `plur_packs_install`, `plur_packs_list`, `plur_capture`, `plur_timeline`, and more) are all reachable via `plur_admin`. Set `PLUR_TOOL_PROFILE=full` to expose all 40 tools directly.
|
|
69
67
|
|
|
70
68
|
## Sync across machines
|
|
71
69
|
|
|
@@ -102,11 +100,9 @@ Default: `~/.plur/`. Everything is plain YAML — open it, read it, edit it.
|
|
|
102
100
|
|
|
103
101
|
## Benchmark
|
|
104
102
|
|
|
105
|
-
|
|
106
|
-
|
|
107
|
-
|
|
108
|
-
| A/B win rate vs no memory | 89% |
|
|
109
|
-
| House rules accuracy | 100% |
|
|
103
|
+
**Retrieval** (LongMemEval R@5): **76.7%** out-of-the-box · **97.0%** with openai-3-large embeddings
|
|
104
|
+
|
|
105
|
+
**Agent task impact:** Haiku + PLUR outperforms Opus *without* memory at ~10× less cost. House rules: **12–0** across Haiku, Sonnet, Opus. A/B win rate: **89%**.
|
|
110
106
|
|
|
111
107
|
[Full methodology →](https://plur.ai/benchmark.html)
|
|
112
108
|
|