@brunobracaioli/aibrain 0.2.0-beta.1 → 0.3.0-beta.14

package/README.md

@@ -2,9 +2,23 @@
 
 Assisted AI Brain context beta client for Linux/WSL with Node.js 22 or newer.
 
-The reviewed native compatibility set is Claude Code 2.1.294, 2.1.295, or
-2.1.296, and Codex CLI 0.161.0, 0.162.0 or 0.162.1. Other versions require a new
-compatibility review before setup.
+The reviewed native compatibility set is Claude Code 2.1.294, 2.1.295, 2.1.296,
+2.1.299, 2.1.300, 2.1.301, 2.1.302 or 2.1.303, and Codex CLI 0.161.0,
+0.162.0, 0.162.1 or 0.163.0. A newer stable Codex 0.x release may continue only
+after its locally generated experimental protocol schemas and transport flags
+match the bundled adapter contract; prereleases, older releases and new majors
+remain rejected. Starting with beta14, newer stable Claude 2.1 patches can
+continue after isolated strict plugin validation, required hook/API checks and
+an enabled-hook test in the native offline runner. Doctor distinguishes
+`host_checked` from `qualified`; this check does not establish full behavioral
+equivalence with a tested native release. Unknown older patches, prereleases
+and other major/minor families remain rejected. No credentials, conversations,
+model requests or JEV calls are used by these compatibility probes.
+
+The Codex terminal bridge permits the native session picker's separate read-only
+connection while keeping the primary conversation and monitor intact. A picker
+cannot compact, start a turn, inject history or take ownership of the primary
+connection. Its failure or timeout closes only that auxiliary connection.
 
 ```sh
 # Install the published beta:
@@ -36,7 +50,7 @@ For users of an earlier unscoped review tarball, close its client sessions,
 run `npm uninstall -g aibrain`, and install this scoped package. Local
 `~/.aibrain` state, notes, and checkpoints are retained.
 
-This `0.2` client adds configurable context profiles. A fresh
+This `0.3` client adds configurable context profiles. A fresh
 setup defaults to `upstream`: 60% trigger, 0.5 keep threshold, six recent items
 protected, and broader completed text-tool eligibility in Claude. Existing
 schema1 configurations keep their previous behavior. After a reviewed v2 backend
@@ -54,5 +68,65 @@ text; they do not claim that a live session is being monitored. Codex keeps nati
 compaction and recovery notes in the same thread. The backend checks paid access
 and unchanged shared quotas before any JEV request.
 
+The `0.3.0-beta.14` Claude launcher uses the durable context strategy and
+records bounded recovery reasons when native summary proof cannot be confirmed:
+JEV selects compatible historical tool evidence, bounded private notes retain
+the selected text, and Claude core summarizes the reduced input. The next
+prompt restores those notes in the same session after verifying native state.
+The final native context contains a summary; this is not summary-free transcript
+replacement. Direct hook-owned history replacement remains experimental because
+resume can replay older content in affected Claude versions.
+
+The notes preserve only the selected retained text that fits the local bound;
+oversized entries are omitted as whole entries and counted. The separate full
+checkpoint has its own retention. Neither path guarantees perfect preservation
+of every internal host field. Shadow prefiltering continues to observe without
+changing tool results. Native qualification and publication evidence is tracked
+in the [release record](https://github.com/brunobracaioli/mcp-ai-brain/blob/main/docs/reference/claude-durable-context-2026-10-14.md).
+
+The launcher grants the native clients the provider-free `brain_context_budget`
+read alongside the existing context tools. It is a short-lived advisory snapshot,
+not a billing meter, reservation, or provider authorization; atomic admission
+remains authoritative.
+
+Codex diagnostics are kept in the launched adapter's private state directory.
+`aibrain diagnostics --client codex --json` shows only bounded stages, closed
+reasons, pseudonymous session identifiers, client versions, and small numeric
+counters. Prompts, notes, paths, credentials, raw errors, and app-server stdout
+are excluded.
+
+Beta12 reads large Codex histories in bounded pages and keeps their complete
+contents in verified private capture segments. It checks history again before
+selection and before native compaction; notes and recovery stay in the same
+thread. Older captures remain readable. A single turn exceeding 8 MiB, a capture
+exceeding the bounded page/item/128 MiB limits, or exhausted private storage is
+reported explicitly. Such a failure does not authorize discarding protected
+history or retrying a paid operation. The private capture directory remains
+bounded to 256 MiB and 128 managed files, with current recovery references
+protected from cleanup.
+
+Claude plan diagnostics distinguish response parsing, identity, expiry and
+history changes. A text JSON plan can use the same 256 KiB limit as structured
+plans; smaller budget/error envelopes retain their separate limits. These local
+diagnostics contain format names and numeric sizes, never plan contents.
+
+Context dashboard telemetry is disabled for fresh and existing configurations.
+Enable it only after reviewing the extended disclosure:
+
+```sh
+aibrain setup --client both --telemetry on --accept-disclosure
+```
+
+Omit `--telemetry` to preserve the stored choice, or use `--telemetry off` to stop
+future-launch reports. Settings are captured when a client starts, so close and
+restart currently monitored sessions after opting out to stop their reporting.
+Enabled reports use the existing authenticated MCP transport and
+contain only a hashed native session key, random lifecycle identifiers, native
+client version, phase, timestamps, numeric counts, and allowed measurements or
+reasons. Prompts, notes, paths, tool names, credentials, and raw errors are not
+reported. The dashboard shows a 30-day visible window; physical cleanup runs on
+the next successful daily cleanup. Delivery failure never fails or triggers a
+context operation.
+
 Read the [canonical subscriber beta guide](https://github.com/brunobracaioli/mcp-ai-brain/blob/main/docs/how-to/aibrain-subscriber-beta.md)
 for setup, support, disclosure, limits, and disable procedures.

package/dist/assets/claude/.claude-plugin/plugin.json

@@ -1,6 +1,6 @@
 {
   "name": "ai-brain-context-compaction",
-  "version": "0.1.0",
+  "version": "0.3.0-beta.14",
   "author": { "name": "AISAC" },
   "description": "AI Brain paid-plan context monitoring with conservative Claude compaction fallback.",
   "hooks": "./hooks/hooks.json",
@@ -89,11 +89,59 @@
       "description": "preserve_internal keeps handled messages intact. upstream_compatible rebuilds selected handled read messages from exposed fields and may lose hidden engine metadata.",
       "default": "preserve_internal"
     },
+    "compactionStrategy": {
+      "type": "string",
+      "title": "Compaction strategy",
+      "description": "direct_experimental returns a hook-owned transcript; native_summary_notes delegates the final summary to Claude core and keeps selected evidence in private notes.",
+      "default": "direct_experimental"
+    },
     "diagnosticMode": {
       "type": "boolean",
       "title": "Emit runtime diagnostics",
       "description": "Opt-in content-free hook entry diagnostics for local pilot verification.",
       "default": false
+    },
+    "telemetryEnabled": {
+      "type": "boolean",
+      "title": "Enable context dashboard telemetry",
+      "description": "Opt-in operational telemetry for the authenticated Contexto dashboard. The launcher enables this only after the extended AI Brain disclosure.",
+      "default": false
+    },
+    "telemetryDisclosed": {
+      "type": "boolean",
+      "title": "Context telemetry disclosure acknowledged",
+      "description": "Permit content-free lifecycle reports containing a hashed session key, bounded measurements, counts, phases and reasons.",
+      "default": false
+    },
+    "telemetryClientVersion": {
+      "type": "string",
+      "title": "Native Claude version",
+      "description": "Launcher-provided native semver used to label telemetry; empty values keep telemetry inactive.",
+      "default": ""
+    },
+    "telemetryNodePath": {
+      "type": "string",
+      "title": "Owned telemetry Node path",
+      "description": "Launcher-provided absolute Node executable path for the report helper.",
+      "default": ""
+    },
+    "telemetryHelperPath": {
+      "type": "string",
+      "title": "Owned telemetry helper path",
+      "description": "Launcher-provided absolute report-only helper path.",
+      "default": ""
+    },
+    "telemetryStateDir": {
+      "type": "string",
+      "title": "Owned telemetry state path",
+      "description": "Launcher-provided private aibrain state directory for the helper.",
+      "default": ""
+    },
+    "telemetryConfigDigest": {
+      "type": "string",
+      "title": "Owned telemetry configuration binding",
+      "description": "Launcher-provided binding for the private endpoint, token and disclosure state.",
+      "default": ""
     }
   }
 }