fix: P3 — normalize variable name drift across core modules

- *memory* → *memory-store* (context, perceive tests, defpackage export)
- *skills-registry* → *skill-registry* (context, reason)
- *logs-lock* → *log-lock*, *system-logs* → *log-buffer* (context → defpackage)
- *cognitive-tools* → *cognitive-tool-registry* (act)
- deterministic-verify → cognitive-verify (act → reason)
These were runtime errors — referenced variables that don't exist.
Harmless until called, but would crash if those code paths were hit.

org/core-context.org

@@ -46,7 +46,7 @@ Filters the Memory store by tag, TODO state, or object type. This is the primary
                  (when tag (unless (search tag (format nil "~a" (getf attrs :TAGS)) :test #'string-equal) (setf match nil)))
                  (when (and todo-state (not (equal state todo-state))) (setf match nil))
                  (when match (push obj results))))
-             *memory*)
+             *memory-store*)
     results))
 #+end_src
 
@@ -82,7 +82,7 @@ Provides a sorted overview of currently loaded system capabilities. Each entry i
     (maphash (lambda (name skill)
                (declare (ignore name))
                (push (list :name (skill-name skill) :priority (skill-priority skill) :dependencies (skill-dependencies skill)) results))
-             *skills-registry*)
+             *skill-registry*)
     (sort results #'> :key (lambda (x) (getf x :priority)))))
 #+end_src
 
@@ -126,9 +126,9 @@ Retrieves the most recent lines from the harness's internal log buffer. The log
 (defun context-logs (&optional limit)
   "Retrieves the most recent lines from the harness's internal log."
   (let ((log-limit (or limit (ignore-errors (parse-integer (uiop:getenv "CONTEXT_LOG_LIMIT"))) 20)))
-    (bt:with-lock-held (*logs-lock*)
-      (let ((count (min log-limit (length *system-logs*)))) 
-        (subseq *system-logs* 0 count)))))
+    (bt:with-lock-held (*log-lock*)
+      (let ((count (min log-limit (length *log-buffer*)))) 
+        (subseq *log-buffer* 0 count)))))
 #+end_src
 
 ** AST to Org Rendering (context-object-render)
@@ -263,7 +263,7 @@ Verifies that the Foveal-Peripheral rendering correctly distinguishes between fo
 (in-suite vision-suite)
 
 (test test-foveal-rendering
-  (clrhash passepartout::*memory*)
+  (clrhash passepartout::*memory-store*)
   (let* ((ast '(:type :HEADLINE :properties (:ID "proj-root" :TITLE "Project" :TAGS ("project"))
                 :contents ((:type :HEADLINE :properties (:ID "node-foveal" :TITLE "Foveal Node")
                              :raw-content "FOVEAL CONTENT" :contents nil)
@@ -276,7 +276,7 @@ Verifies that the Foveal-Peripheral rendering correctly distinguishes between fo
       (is (not (search "PERIPHERAL CONTENT" output))))))
 
 (test test-awareness-budget
-  (clrhash passepartout::*memory*)
+  (clrhash passepartout::*memory-store*)
   (ingest-ast '(:type :HEADLINE :properties (:ID "p1" :TITLE "Project 1" :TAGS ("project")) :contents nil))
   (ingest-ast '(:type :HEADLINE :properties (:ID "p2" :TITLE "Project 2" :TAGS ("project")) :contents nil))
   (let ((output (context-awareness-assemble)))

org/core-defpackage.org

@@ -54,7 +54,7 @@ The package definition. All public symbols are exported here.
    #:lookup-object
    #:list-objects-by-type
    #:org-id-new
-   #:*memory*
+   #:*memory-store*
    #:*history-store*
    #:org-object
    #:make-org-object

org/core-loop-act.org

@@ -18,7 +18,7 @@ The key architectural choice: **actuators are not privileged**. The same dispatc
 
 ** Why Dispatch-Action Verifies Again?
 
-The Reason stage already ran every proposed action through the deterministic engine. So why does ~loop-gate-act~ call ~deterministic-verify~ again?
+The Reason stage already ran every proposed action through the deterministic engine. So why does ~loop-gate-act~ call ~cognitive-verify~ again?
 
 Because a skill's deterministic gate runs during Reason, but between Reason and Act, the action might have been transformed by the pipeline (metadata added, format normalized). The last-mile verification catches any transformation that might have introduced an unsafe property. It's the same philosophy as "trust but verify" — the second check is cheap and catches a class of bugs that would otherwise be silent data corruption.
 
@@ -135,7 +135,7 @@ The tool's return value is packed into a ~:tool-output~ event and fed back into
          (depth (getf context :depth 0))
          (meta (getf context :meta))
          (source (getf meta :source))
-         (tool (gethash (string-downcase (string tool-name)) *cognitive-tools*)))
+         (tool (gethash (string-downcase (string tool-name)) *cognitive-tool-registry*)))
     (if tool
         (handler-case
             (let* ((clean-args (if (and (listp tool-args) (listp (car tool-args))) (car tool-args) tool-args))
@@ -189,7 +189,7 @@ After dispatch, the gate captures any feedback produced by the actuation (tool o
          (feedback nil))
     (when approved
       (let* ((original-type (getf approved :type))
-             (verified (deterministic-verify approved signal)))
+             (verified (cognitive-verify approved signal)))
         (if (and (listp verified) (member (getf verified :type) '(:LOG :EVENT)) (not (member original-type '(:LOG :EVENT))))
             (progn
               (log-message "ACT BLOCKED: Action failed last-mile deterministic check.")

org/core-loop-perceive.org

@@ -157,11 +157,11 @@ Verifies that the perceive gate correctly ingests AST nodes into memory and that
 (in-suite pipeline-perceive-suite)
 
 (test test-loop-gate-perceive
-  (clrhash passepartout::*memory*)
+  (clrhash passepartout::*memory-store*)
   (let* ((signal (list :type :EVENT :payload (list :sensor :buffer-update :ast (list :type :HEADLINE :properties (list :ID "test-node" :TITLE "Test") :contents nil))))
          (result (loop-gate-perceive signal)))
     (is (eq :perceived (getf result :status)))
-    (is (not (null (gethash "test-node" passepartout::*memory*))))))
+    (is (not (null (gethash "test-node" passepartout::*memory-store*))))))
 
 (test test-depth-limiting
   (let ((runaway-signal (list :type :EVENT :depth 11 :payload (list :sensor :heartbeat))))

org/core-loop-reason.org

@@ -194,7 +194,7 @@ The system prompt assembly order — identity, tools, context, logs, mandates
                                           (let ((aug-text (ignore-errors (funcall aug-fn context))))
                                             (when (and aug-text (stringp aug-text) (> (length aug-text) 0))
                                               (setf augments (concatenate 'string augments aug-text (string #\Newline))))))))
-                                    *skills-registry*)
+                                    *skill-registry*)
                            (when (> (length augments) 0) augments)))
          (system-prompt (format nil "IDENTITY: ~a~a~%~%TOOLS:~%~a~%~%CONTEXT:~%~a~%~%LOGS:~%~a~%~a" 
                                 assistant-name reflection-feedback tool-belt global-context system-logs
@@ -231,7 +231,7 @@ This architecture makes safety compositional: each skill adds one constraint. Th
                (declare (ignore name))
                (when (skill-deterministic-fn skill)
                  (push skill skills)))
-             *skills-registry*)
+             *skill-registry*)
     (setf skills (sort skills #'> :key #'skill-priority))
     (dolist (skill skills)
       (let ((trigger (skill-trigger-fn skill))
@@ -302,7 +302,7 @@ Verifies that the deterministic engine correctly rejects unsafe actions (like ~r
 (in-suite pipeline-reason-suite)
 
 (test test-decide-gate-safety
-  (clrhash passepartout::*skills-registry*)
+  (clrhash passepartout::*skill-registry*)
   (passepartout::defskill :mock-safety
     :priority 50
     :trigger (lambda (ctx) (declare (ignore ctx)) t)