wicked-garden 12.52.1 → 12.53.0

package/.claude-plugin/marketplace.json

@@ -1 +1 @@
-{"name": "wicked-garden", "owner": {"name": "Mike Parcewski"}, "plugins": [{"name": "wicked-garden", "source": "./", "description": "Your coding agent already plans and swarms \u2014 Wicked Garden is the curated toolkit for what it can't: proof-not-claims gates, code links grep can't see, deterministic refactor, cross-session memory, multi-model second opinions (fallback seats labelled). Don't fight the harness; fill its gaps. The gate needs wicked-vault (a self-contained infra peer, installed directly) and re-derives through garden's built-in engine; estate/bus are opt-in layers (wicked-testing and wicked-brain are retired \u2014 QE ships in-catalog as the qe domain, memory lives in wicked-estate).", "version": "12.52.1"}], "version": "12.52.1"}
+{"name": "wicked-garden", "owner": {"name": "Mike Parcewski"}, "plugins": [{"name": "wicked-garden", "source": "./", "description": "Your coding agent already plans and swarms \u2014 Wicked Garden is the curated toolkit for what it can't: proof-not-claims gates, code links grep can't see, deterministic refactor, cross-session memory, multi-model second opinions (fallback seats labelled). Don't fight the harness; fill its gaps. The gate needs wicked-vault (a self-contained infra peer, installed directly) and re-derives through garden's built-in engine; estate/bus are opt-in layers (wicked-testing and wicked-brain are retired \u2014 QE ships in-catalog as the qe domain, memory lives in wicked-estate).", "version": "12.53.0"}], "version": "12.53.0"}

package/.claude-plugin/plugin.json

@@ -1,6 +1,6 @@
 {
   "name": "wicked-garden",
-  "version": "12.52.1",
+  "version": "12.53.0",
   "description": "Your coding agent's harness already plans and swarms; Wicked Garden is the curated toolkit for what it can't do alone. Proof, not claims \u2014 gates re-derive 'done' through the internal loom engine (fail-closed, independently attested via wicked-vault), so a self-asserted pass can't lie its way green. Relationships grep can't see \u2014 the wicked-estate code graph + injected bus/dispatch/capability/archetype edges power blast-radius & lineage. Plus deterministic multi-file refactor (wicked-patch), cross-session memory + knowledge (the mem domain over wicked-estate), multi-model second opinions (fallback seats labelled) (jam:council), portable expertise as on-demand skill-refs, and evidence-gated testing (the in-catalog qe domain). It reads the shape of your work to apply the right rigor \u2014 steering, not a pipeline \u2014 then gets out of the way. The compiler stamps a self-contained evidence gate into any repo that runs with no wicked-garden installed. Don't fight the harness; fill its gaps. The evidence gate requires wicked-vault, a self-contained infra peer wicked-garden installs directly; wicked-bus is an opt-in layer \u2014 install what you'll use, the toolkit works without the rest.",
   "wicked_vault_version": "^0.5.0",
   "wicked_ledger_version": "^0.6.0",

package/CHANGELOG.md

@@ -2,6 +2,11 @@
 
 ## [Unreleased]
 
+## [12.53.0] — 2026-10-11
+
+### Changed
+- **Learned code facts are filed as knowledge, not memory (PR #1285).** repo-learn submits code facts as knowledge proposals, and the shim lifts its refusal of `kind_type: "knowledge"` and requires each proposal's `source`. Needs wicked-estate 0.29.2+ (estate K3, the knowledge-proposal type); an older estate refuses the proposal.
+
 ## [12.52.1] — 2026-10-11
 
 ### Changed

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "wicked-garden",
-  "version": "12.52.1",
+  "version": "12.53.0",
   "type": "module",
   "description": "Curated toolkit for what AI coding agents can't do alone: evidence-gated work, codegraph relationships, deterministic multi-file refactor, multi-model second opinions, and cross-session memory. Installs as a Claude Code plugin.",
   "keywords": [

package/scripts/_estate_client.py

@@ -1102,13 +1102,10 @@ def propose(
     run's `WICKED_RUN_PROJECT` supplies): an unscoped proposal is refused here
     with the shape-error code, so it never lands project-less in the queue.
     """
-    if kind_type == "knowledge":
-        # Gate (DES-memory-vs-knowledge K3): estate's approve promotes only memory + policy:*,
-        # so a knowledge proposal would strand in the queue. Lift when estate ships the type.
+    if kind_type == "knowledge" and not (isinstance(payload, dict) and str(payload.get("source") or "").strip()):
+        # DES-memory-vs-knowledge K3 (estate 0.29.2): knowledge is acquired, so it names its source.
         return {"ok": False, "code": -32602,
-                "reason": 'kind_type "knowledge" is not approvable yet (estate promotes only '
-                          'memory and policy:*): keep the object in your report under '
-                          '"knowledge (pending)"; do not re-file it as a memory'}
+                "reason": 'a knowledge proposal needs payload.source (where it was read, e.g. "<repo>:<path>")'}
     facets = _scope_facets_to_run_project(facets)
     project = (facets or {}).get("project")
     if not isinstance(project, str) or not project.strip():

package/skills/repo-learn/SKILL.md

@@ -138,11 +138,9 @@ when the proposal is approved, so the learning is lost):
 - **Knowledge** — how the code works, read from the code (*acquired*, so it is
   knowledge, not memory: DES-memory-vs-knowledge). `kind_type: "knowledge"`; payload
   `{"content": "<1–3 sentences>", "class": "concept", "source": "<repo>:<path>",
-  "source_rev": "<commit sha>"}`. Most repo-learn findings are this kind. **Not
-  submittable yet:** estate cannot approve a knowledge proposal until its knowledge
-  proposal type ships, so the shim refuses `kind_type: "knowledge"`. List these objects
-  in your report's fenced `json` block under `knowledge (pending)` and leave them out
-  of the `wicked-capture-report` counts. Never propose a graph metric (PageRank rank,
+  "source_rev": "<commit sha>"}`. Most repo-learn findings are this kind. Submit it
+  like any proposal (estate 0.29.2+ approves it into the knowledge store); count it in
+  the `wicked-capture-report` like the rest. Never propose a graph metric (PageRank rank,
   churn count, blast-radius size): estate's tools re-derive those on demand.
 - **Memory** — only something *witnessed* while studying (a gotcha you hit, with
   why) or *told* by the operator. Split a mixed finding: "we chose fail-closed

package/skills/repo-learn/refs/capture-proposals.md

@@ -73,12 +73,10 @@ estate/steering types — pick the one the rule governs, not a freeform tag.
 What you learn by reading code is **acquired**, so it is knowledge, not memory
 (DES-memory-vs-knowledge §3).
 
-**Not submittable yet.** Estate's `proposal.approve` promotes only `memory` and
-`policy:*`; a knowledge proposal would sit unapprovable in the queue, so the shim
-refuses `kind_type: "knowledge"` (`-32602`) until estate's knowledge proposal type
-ships. Until then, put each object below in your report's fenced `json` block under
-`knowledge (pending)`, and leave them out of the `wicked-capture-report` counts. Do
-not re-file a code fact as a `memory` to get it through. The shape is:
+Submit it through the shim's `propose` like a memory. Estate 0.29.2+ approves a
+knowledge proposal into the knowledge store, keeping its `source`, `observed_at` and
+`source_rev`. The shim refuses one with no `source` (`-32602`). Never re-file a code fact as a
+`memory`. The shape is:
 
 ```json
 {