panoma 0.11.1 → 0.13.0

package/README.md

@@ -16,7 +16,7 @@ No install, no account, no network: it analyzes the folder and prints what proje
 - `npx panoma scan <path> -d` — copies of the same project, grouped.
 - `npx panoma md check <path>` — checks the project's `AGENTS.md` / `CLAUDE.md`: paths that no longer exist, scripts that are gone, versions that no longer run.
 - `npx panoma md init <path>` — writes your agents a truthful starting point (needs the catalog up).
-- `npx panoma handoff --to codex` — continues the newest conversation in this folder in another agent: written into that agent's own history so its normal resume finds it, the original untouched. Add `--dry-run` to see what would travel first.
+- `npx panoma handoff --to codex` — continues the newest conversation in this folder in another agent: written into that agent's own history so its normal resume finds it, the original untouched. Add `--dry-run` to see what would travel first, with what each tier weighs in tokens; `--tier compact` sends the digest and the newest turns, and `--digest model` lets the connected model write that digest from the whole conversation, in windows, one counted call each.
 
 ## The app
 
@@ -48,7 +48,8 @@ npx panoma memory allow claude-code capture --all --notice 2
 npx panoma memory status
 ```
 
-More at [panoma.ai](https://panoma.ai).
+More at [panoma.ai](https://panoma.ai); the memory told step by step, plainly or technically, at
+[panoma.ai/memory](https://www.panoma.ai/memory).
 
 ## Licence
 

package/app/BUILD-INFO.json

@@ -1,6 +1,6 @@
 {
-  "version": "0.11.1",
-  "commit": "3ae0a4bee2cd4d5ce47806797259e9aa885d7c96",
+  "version": "0.13.0",
+  "commit": "eebdc7208b7b78fe04662af9c3aa18a83f36e4ce",
   "arbolLimpio": true,
   "lockfile": "54405bd4e3a7fb6880050331687a1be4e05235a0ffe1b4d3dead3424d051433d",
   "node": "v26.7.0",

package/app/apps/web/.next-bundle/BUILD_ID

@@ -1 +1 @@
-z5NZxxbhRK5pbTErTEcgK
+G80IkXH9NgWvVgFFqBtOe