@plurnk/plurnk-grammar 0.54.0 → 0.56.0

package/README.md

@@ -21,7 +21,7 @@ const result = PlurnkParser.parse(input);
 
 Discriminate on `item.kind`. For `statement` items, narrow on `statement.op` (one of `FIND READ EDIT COPY MOVE OPEN FOLD SEND EXEC KILL PLAN`) to access per-OP typed fields. Full API: [SPEC.md §12](SPEC.md#12-public-api).
 
-`parsePath(raw)` is a top-level helper that decomposes a path/URI string into a `ParsedPath` (the same decomposition applied to `(target)` slots). Reach for it to resolve a **COPY destination**: COPY's body is an opaque string — a destination URI for an entry copy, a prompt for a run fork (`run:///`) — so the consumer interprets it by scheme and calls `parsePath` for the destination case. (MOVE destinations arrive pre-parsed; COPY's don't, because its body is polymorphic.)
+`parsePath(raw)` is a top-level helper that decomposes a path/URI string into a `ParsedPath` (the same decomposition applied to `(target)` slots). Reach for it to resolve a **COPY destination**: COPY's body is an opaque string — a destination URI for an entry copy, a prompt for a run fork (`run://`) — so the consumer interprets it by scheme and calls `parsePath` for the destination case. (MOVE destinations arrive pre-parsed; COPY's don't, because its body is polymorphic.)
 
 ## cli
 

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@plurnk/plurnk-grammar",
-  "version": "0.54.0",
+  "version": "0.56.0",
   "description": "ANTLR4 grammar for the Plurnk LLM agent protocol",
   "type": "module",
   "license": "MIT",

package/plurnk.md

@@ -34,10 +34,10 @@ EXEC defaults to `sh`; override with an optional executor (`sqlite`, `node`, etc
 
 The agent maintains two surfaces for budgeting working-memory tokens:
 
-- **Log** — the record of every operation. FOLD contracts a log row to its one-line summary and saves tokens; OPEN shows the complete record but spends from your `tokensFree` context tokens. Non-destructive — folded rows remain listed and re-OPENable.
+- **Log** — the record of every operation. FOLD contracts a log row to its one-line summary and saves tokens; OPEN shows the complete record but spends from your `tokensFree` context tokens. Non-destructive — FOLDed rows remain listed and re-OPENable.
 - **`plurnk:///manifest.json`** — what's available: the complete unranked directory of every entry. Query it to discover available entries.
 
-OPEN and FOLD operate on the log only.
+OPEN and FOLD operate on the log only. Log items are read-only, but can be KILLed.
 
 ## `<Line> / <Result>`
 
@@ -122,8 +122,8 @@ Body content is character-perfect, exactly matching whitespace.
 <<MOVE[final](known:///draft/answer.md):known:///final/answer.md:MOVE
 <<KILL(obsolete/file.md)::KILL
 <<KILL(sh:///3/1/2)::KILL
-<<EDIT(run:///capital-check):Find the capital of France.:EDIT
-<<COPY(run:///.):Re-derive the capital from a primary source.:COPY
+<<EDIT(run://capital-checker):Find the capital of France.:EDIT
+<<COPY(run://.):Re-derive the capital from a primary source.:COPY
 <<OPEN(log:///**/get)<1,10>::OPEN
 <<FIND(known:///**)<5>:~french revolutionary history:FIND
 <<FIND(known:///**)<0.7>:~french territorial concessions:FIND
@@ -135,7 +135,7 @@ Body content is character-perfect, exactly matching whitespace.
 <<FIND(src/**):@<createCoder:FIND
 
 <<EDIT[tutorial,training,scripts](example.sh):#!/usr/bin/env sh
-echo "Hello, world!" > hello.txt
+echo "Taxonomic path names and folksonomic tags on known entries improve reasoning and recall!" > advice.txt
 :EDIT
 
 <<EXEC:
@@ -145,7 +145,7 @@ chmod +x ./example.sh
 
 <<EXEC[sqlite]:SELECT 22.0 / 7.0;:EXEC
 
-<<SEND(run:///capital-check):{"hint":"prefer primary sources over wikis"}:SEND
+<<SEND(run://capital-checker):{"hint":"known entries are your persistent memory"}:SEND
 
-<<SEND[102]:decomposed prompt into unknowns; plan initialized:SEND
+<<SEND[102]:Forking a research run, optimizing log relevance...:SEND
 <<SEND[200]:Paris:SEND