Garry's Opinionated OpenClaw/Hermes Agent Brain
TypeScript
v0.60.575.0 chore(release): restamp onto master v0.60.574.0
garrytan9c7f59871919
diff --git a/.claude-plugin/plugin.json b/.claude-plugin/plugin.jsonindex b282022060..1e50f3c6ff 100644--- a/.claude-plugin/plugin.json+++ b/.claude-plugin/plugin.json@@ -1,6 +1,6 @@{"name": "gbrain",- "version": "0.60.574.0",+ "version": "0.60.575.0","description": "Personal knowledge brain for your coding agent — hybrid search, synthesis, graph traversal, and durable cross-session memory over Postgres/PGLite with pgvector, plus a curated brain-first skill set.","author": {"name": "Garry Tan",diff --git a/.codex-plugin/plugin.json b/.codex-plugin/plugin.jsonindex 2578a1a764..22095a48c0 100644--- a/.codex-plugin/plugin.json+++ b/.codex-plugin/plugin.json@@ -1,6 +1,6 @@{"name": "gbrain",- "version": "0.60.574.0",+ "version": "0.60.575.0","description": "Personal knowledge brain for your coding agent — hybrid search, synthesis, graph traversal, and durable cross-session memory over Postgres/PGLite with pgvector, plus a curated brain-first skill set.","author": {"name": "Garry Tan",diff --git a/BOOTSTRAP_FOR_AGENTS.md b/BOOTSTRAP_FOR_AGENTS.mdindex bf43f45797..6219218c73 100644--- a/BOOTSTRAP_FOR_AGENTS.md+++ b/BOOTSTRAP_FOR_AGENTS.md@@ -1,4 +1,4 @@-<!-- gbrain-runbook-stamp: 0.60.574.0 -->+<!-- gbrain-runbook-stamp: 0.60.575.0 --><!-- This stamp must equal the VERSION file at every release; CI enforces it(scripts/check-bootstrap-tag.sh). `gbrain bootstrap status` compares it tothe installed binary and warns on skew. -->diff --git a/openclaw.plugin.json b/openclaw.plugin.jsonindex 2d133e5455..066c258f1c 100644--- a/openclaw.plugin.json+++ b/openclaw.plugin.json@@ -1,7 +1,7 @@{"id": "gbrain-context-engine","name": "gbrain",- "version": "0.60.574.0",+ "version": "0.60.575.0","description": "Personal knowledge brain with Postgres + pgvector hybrid search","family": "bundle-plugin","kind": "context-engine",diff --git a/plugin-variants/gbrain-coding/.claude-plugin/plugin.json b/plugin-variants/gbrain-coding/.claude-plugin/plugin.jsonindex d28bb4bd65..7130dbff25 100644--- a/plugin-variants/gbrain-coding/.claude-plugin/plugin.json+++ b/plugin-variants/gbrain-coding/.claude-plugin/plugin.json@@ -1,6 +1,6 @@{"name": "gbrain-coding",- "version": "0.60.574.0",+ "version": "0.60.575.0","description": "Brain-first coding agent working inside a repo: retrieval, routing, ingest discipline, correction hygiene. Default persona for the claude-code harness bridge; also published as the gbrain-coding marketplace variant. (persona variant of the gbrain plugin — 21 skills)","author": {"name": "Garry Tan",diff --git a/plugin-variants/gbrain-coding/.codex-plugin/plugin.json b/plugin-variants/gbrain-coding/.codex-plugin/plugin.jsonindex 3144f40154..2b1146e272 100644--- a/plugin-variants/gbrain-coding/.codex-plugin/plugin.json+++ b/plugin-variants/gbrain-coding/.codex-plugin/plugin.json@@ -1,6 +1,6 @@{"name": "gbrain-coding",- "version": "0.60.574.0",+ "version": "0.60.575.0","description": "Brain-first coding agent working inside a repo: retrieval, routing, ingest discipline, correction hygiene. Default persona for the claude-code harness bridge; also published as the gbrain-coding marketplace variant. (persona variant of the gbrain plugin — 21 skills)","author": {"name": "Garry Tan",diff --git a/plugin-variants/gbrain-coding/README.md b/plugin-variants/gbrain-coding/README.mdindex 1858fb6aa6..d1be682c40 100644--- a/plugin-variants/gbrain-coding/README.md+++ b/plugin-variants/gbrain-coding/README.md@@ -1,4 +1,4 @@-<!-- gbrain-plugin-tree-stamp: 0.60.574.0 -->+<!-- gbrain-plugin-tree-stamp: 0.60.575.0 --># gbrain-coding (generated persona variant — do not hand-edit)Brain-first coding agent working inside a repo: retrieval, routing, ingest discipline, correction hygiene. Default persona for the claude-code harness bridge; also published as the gbrain-coding marketplace variant.diff --git a/plugin-variants/gbrain-daily/.claude-plugin/plugin.json b/plugin-variants/gbrain-daily/.claude-plugin/plugin.jsonindex 5fbcfaedf0..b46b1b85ca 100644--- a/plugin-variants/gbrain-daily/.claude-plugin/plugin.json+++ b/plugin-variants/gbrain-daily/.claude-plugin/plugin.json@@ -1,6 +1,6 @@{"name": "gbrain-daily",- "version": "0.60.574.0",+ "version": "0.60.575.0","description": "Personal knowledge-brain daily use: meetings, tasks, briefings, reading, research. Published as the gbrain-daily marketplace variant. (persona variant of the gbrain plugin — 20 skills)","author": {"name": "Garry Tan",diff --git a/plugin-variants/gbrain-daily/.codex-plugin/plugin.json b/plugin-variants/gbrain-daily/.codex-plugin/plugin.jsonindex 70660336af..b64530a739 100644--- a/plugin-variants/gbrain-daily/.codex-plugin/plugin.json+++ b/plugin-variants/gbrain-daily/.codex-plugin/plugin.json@@ -1,6 +1,6 @@{"name": "gbrain-daily",- "version": "0.60.574.0",+ "version": "0.60.575.0","description": "Personal knowledge-brain daily use: meetings, tasks, briefings, reading, research. Published as the gbrain-daily marketplace variant. (persona variant of the gbrain plugin — 20 skills)","author": {"name": "Garry Tan",diff --git a/plugin-variants/gbrain-daily/README.md b/plugin-variants/gbrain-daily/README.mdindex ea56f7847b..a0b850bbf4 100644--- a/plugin-variants/gbrain-daily/README.md+++ b/plugin-variants/gbrain-daily/README.md@@ -1,4 +1,4 @@-<!-- gbrain-plugin-tree-stamp: 0.60.574.0 -->+<!-- gbrain-plugin-tree-stamp: 0.60.575.0 --># gbrain-daily (generated persona variant — do not hand-edit)Personal knowledge-brain daily use: meetings, tasks, briefings, reading, research. Published as the gbrain-daily marketplace variant.diff --git a/plugin/README.md b/plugin/README.mdindex 4fb8715c23..490d99ee68 100644--- a/plugin/README.md+++ b/plugin/README.md@@ -1,4 +1,4 @@-<!-- gbrain-plugin-tree-stamp: 0.60.574.0 -->+<!-- gbrain-plugin-tree-stamp: 0.60.575.0 --># gbrain plugin skill tree (generated — do not hand-edit)This tree is the curated skill set for the gbrain Codex and Claude Codediff --git a/src/core/behavior-change-notice.ts b/src/core/behavior-change-notice.tsindex 23a57f48c7..2451a4bc99 100644--- a/src/core/behavior-change-notice.ts+++ b/src/core/behavior-change-notice.ts@@ -1063,7 +1063,7 @@ export const BEHAVIOR_CHANGES: ReadonlyArray<{ since: string; text: ChangeText;{ since: '0.60.337.0', text: 'Sync now imports a folder-scaffolding file below a source\'s root (`index.md`, `README.md`, `log.md`, `schema.md` or `RESOLVER.md`) as a page when its frontmatter has `type:` or `title:`, for example `projects/acme-example/index.md`. A bare one at any depth and every one at the source root stay skipped. The first sync after the upgrade imports the existing typed files once and deletes or rewrites nothing else; deleting such a file later soft-deletes its page. `gbrain config set sync.metafile_pages off` restores the old skip. Default change approved by Garry, "take all recommendations", 2026-10-20.' },// GBRA-156 retrieval wave.{ since: '0.60.338.0', text: 'On Postgres, a keyword search whose words match more than 50,000 chunks now ranks the first 50,000 matches that pass its filters instead of all of them, so a common-word query on a large brain returns ranked rows in about a second instead of running into the keyword statement timeout and returning none; hybrid search reports `keyword_candidates_incomplete` when the bound applies. Its any-word retry (after an all-words match finds nothing) now matches only the words the planner statistics say fit about 20,000 chunks, still ranking with every word, and also reports `keyword_candidates_incomplete` when it drops a word.' },- { since: 'TBD', text: 'New opt-in keyword phrase factor (#7175), off by default: `gbrain config set search.keyword_phrase true` makes keyword ranking (the keyword arm of search and query, and `match: keyword`) multiply a chunk that holds the query as one contiguous phrase (stemmed, as `phraseto_tsquery` reads it) by 2, so a rare exact phrase in a long transcript ranks above short pages that only repeat its words. A one-word query is unchanged. Default ranking does not change.' },+ { since: '0.60.575.0', text: 'New opt-in keyword phrase factor (#7175), off by default: `gbrain config set search.keyword_phrase true` makes keyword ranking (the keyword arm of search and query, and `match: keyword`) multiply a chunk that holds the query as one contiguous phrase (stemmed, as `phraseto_tsquery` reads it) by 2, so a rare exact phrase in a long transcript ranks above short pages that only repeat its words. A one-word query is unchanged. Default ranking does not change.' },{ since: '0.60.338.0', text: 'The keyword arm keeps one chunk per page in its candidate pool (Postgres and PGLite), so a page with hundreds of matching chunks no longer pushes every other matching page out of the results.' },{ since: '0.60.338.0', text: 'The title arm\'s any-word retry (after an all-words title match finds nothing) now matches page titles only, as remote callers already did, and drops repeated words (#7135): a date in a question no longer pulls every dated timeline into ranking, which cut a 3-5 s title arm to well under a second on a large brain. A local all-words title match is unchanged.' },// PR 4 E (one provider call per chunk content version, gbrain-evals#180). `since` is stamped at the merge slot.@@ -1427,8 +1427,8 @@ export const BEHAVIOR_CHANGES: ReadonlyArray<{ since: string; text: ChangeText;{ since: '0.60.570.0', id: 'get_page_selector_precedence_gbra280', text: 'get_page given several window selectors now reads one window instead of refusing with invalid_params: cursor, then around_chunk_id, then lines, then bytes, and a cursor that cannot read the page or an around_chunk_id no longer on it gives way to the next. body full, include_content and content_only beside a selector are set aside, so the read stays a window and never returns the whole page. The response\'s window_selector names what applied and what was ignored; a call with one selector is unchanged.' },{ since: '0.60.574.0', id: 'embedded_scope_required_gbra283', text: 'Experimental embedded engines (`openEngine` with `embedded_core: true`) only; the CLI, `gbrain serve` and mounts are unchanged. While such an engine is open, an embed, chat, expansion or rerank call that entered no scope now refuses `engine_scope_required` (new code) even when the host configured a process gateway with `configureGateway`; before, that configuration let unscoped calls use the process keys, and the refusal was `engine_scope_missing`, which now only means a provider the engine lacks or a closed engine. Wrap such calls in `withEngineScope(engine, fn)`, or `withLocalScope(fn)` for calls that belong to the process gateway (both from `gbrain/engine-open`); `hybridSearch`, the import paths, minion jobs and `POST /mcp` enter their engine\'s scope themselves. `openEngine` also refuses an unknown settings key, top-level or nested, with `engine_settings_invalid` naming the key, where it used to ignore it. An embedded engine\'s chat usage is no longer on the process usage stack and survives `reconnect()`. See docs/ENGINES.md "Upgrading an embedded host".' },// GBRA-156 (#7491, #7492). `since` is stamped at the merge slot.- { since: 'TBD', id: 'keyword_identifier_probe', text: 'A search or query naming an identifier (`#412`, a 6+ digit run, a code like INV-2026-00417, an email or a URL) together with other words now runs one more keyword statement: the arm\'s own statement, with the same scope, visibility and filters, matching every identifier the query names. Those rows vote like strict keyword rows, so the page holding the identifier is no longer dropped when filler words ("What did the latest weekly digest (#412) say?") keep it from matching every word, and a title holding every identifier leads the title arm\'s any-word retry. The keyword arm\'s any-word retry ranks rare words first for questions of up to 12 words (it was 4). Queries without an identifier keep their keyword rows. `_meta.retrieval.coverage` counts a pending page only when a query word is a whole token of its title or slug (`412` no longer matches `t004120`), counts pending image pages only when an image arm ran, and adds `could_affect.pending_images`; the `projection_pending` notice says text pages are being indexed and that image pages wait for `gbrain projections drain` or a paid re-import, with doctor\'s `text_projection_readiness` as the fix.' },- { since: 'TBD', id: 'keyword_slug_handles', text: 'Keyword search resolves slug handles: a hyphenated token with a letter in it (`dan-example`), bare or quoted on its own, now also matches the pages whose slug ends in exactly that handle (`people/dan-example`), under the same scope, filters and visibility as any keyword match. Those pages lead `match: "keyword"` results and its count, and lead the hybrid keyword arm as rows that vote like strict rows, so `"dan-example" OR "erin-example"` with `--types \'["person"]\'` returns the two person pages instead of nothing. Queries without such a token run as before.' },+ { since: '0.60.575.0', id: 'keyword_identifier_probe', text: 'A search or query naming an identifier (`#412`, a 6+ digit run, a code like INV-2026-00417, an email or a URL) together with other words now runs one more keyword statement: the arm\'s own statement, with the same scope, visibility and filters, matching every identifier the query names. Those rows vote like strict keyword rows, so the page holding the identifier is no longer dropped when filler words ("What did the latest weekly digest (#412) say?") keep it from matching every word, and a title holding every identifier leads the title arm\'s any-word retry. The keyword arm\'s any-word retry ranks rare words first for questions of up to 12 words (it was 4). Queries without an identifier keep their keyword rows. `_meta.retrieval.coverage` counts a pending page only when a query word is a whole token of its title or slug (`412` no longer matches `t004120`), counts pending image pages only when an image arm ran, and adds `could_affect.pending_images`; the `projection_pending` notice says text pages are being indexed and that image pages wait for `gbrain projections drain` or a paid re-import, with doctor\'s `text_projection_readiness` as the fix.' },+ { since: '0.60.575.0', id: 'keyword_slug_handles', text: 'Keyword search resolves slug handles: a hyphenated token with a letter in it (`dan-example`), bare or quoted on its own, now also matches the pages whose slug ends in exactly that handle (`people/dan-example`), under the same scope, filters and visibility as any keyword match. Those pages lead `match: "keyword"` results and its count, and lead the hybrid keyword arm as rows that vote like strict rows, so `"dan-example" OR "erin-example"` with `--types \'["person"]\'` returns the two person pages instead of nothing. Queries without such a token run as before.' },];/** The newest disclosed change's release: the notice id moves only when a release adds rows. */diff --git a/templates/bootstrap/template-repo/README.md b/templates/bootstrap/template-repo/README.mdindex 6d0c469548..42637300f7 100644--- a/templates/bootstrap/template-repo/README.md+++ b/templates/bootstrap/template-repo/README.md@@ -1,6 +1,6 @@# gbrain agent workspace — template-<!-- gbrain-template-stamp: 0.60.574.0 -->+<!-- gbrain-template-stamp: 0.60.575.0 -->This repository is the **"Use this template"** distribution artifact for a[gbrain](https://github.com/garrytan/gbrain) personal-agent workspace — the same