@gmickel/gno 2.7.1 → 2.8.1

package/README.md

@@ -139,7 +139,7 @@ See the [guide](docs/COMPILED-CONTEXT.md).
 
 <!-- public-truth:current-version -->
 
-> Current source version: **v2.7.1**. See [CHANGELOG.md](./CHANGELOG.md).
+> Current source version: **v2.8.1**. See [CHANGELOG.md](./CHANGELOG.md).
 
 <!-- /public-truth -->
 
@@ -216,7 +216,8 @@ See the [guide](docs/COMPILED-CONTEXT.md).
   and traces. Authentication never overrides policy.
 - **Source availability (`any` | `local`)**: opt-in `local` indexes only files
   already on disk and never makes a cloud provider download one. Supported on
-  macOS with Google Drive, iCloud Drive, and OneDrive SharePoint library roots.
+  macOS with Google Drive (My Drive and Shared drives), iCloud Drive, and
+  OneDrive SharePoint library roots.
   Cloud-only files are skipped and reported, not treated as conversion errors,
   and documents under a cloud-only folder stay indexed. Other platforms fail
   with an error. On a 5,000-file all-local collection, `local` scans about 1%

package/assets/skill/SKILL.md

@@ -352,6 +352,15 @@ offline audits inspect parsed local links, explicitly declared capture/logical-
 record provenance, and observable source/index freshness. They never repair,
 rewrite, persist findings, judge factual truth, or replace retrieval.
 
+`--max-findings` accepts `all` (MCP `maxFindings: "all"`) to export every
+finding. Link findings carry `referenceKind`, `resolutionStatus`, and
+`resolvedScope` in their evidence detail; ambiguous vault links list the tied
+`candidates`. `links.outside-index` `info` findings are link targets that
+exist in the vault but are not indexed (attachments, excluded or unindexed
+folders); Obsidian resolves them, so do not report them as broken links.
+Report `truncation.snapshotTruncated` as "totals cover the bounded snapshot,
+not the whole index".
+
 Treat exit `4` as a complete report with findings. Exit `5` or report status
 `partial`/`changed_during_audit` means evidence is unavailable, inconclusive,
 cancelled, truncated, or repeatedly changed—never healthy. Preserve stable
@@ -389,7 +398,7 @@ remaining steps apply under the default `full` profile.
    - MCP text is the compact `gno-context-agent-v1` evidence projection. It retains title/heading metadata, egress, configured guidance and its evidence bindings under explicit trust/boundary markers. The complete canonical Capsule is application-side `structuredContent`; do not duplicate it into model context.
 3. Use `gno_ask` only for explicit local verified synthesis. Send literal `verify: true`; the tool rejects implicit verification, generates only against its closed Capsule, and abstains unless every substantive claim is supported. Preserve exact spans, gaps, semantic capability state, and abstention. This does not guarantee corpus completeness or source truth.
 4. Use `gno_query` for interactive lookup or manual retrieval control. It returns snippets plus `uri`, `docid`, often `line`, and sometimes `context`. Treat `context` as user-configured guidance for interpreting that exact result; cite source content at the returned URI/lines, not the guidance itself. Bounded graph expansion is on by default; set `graph: false` or `noGraph: true` only for an explicit BM25/vector-only path.
-5. Use graph/link expansion for relationship context: `gno_graph_query` for typed relationship traversal, `gno_graph_neighbors` for nearby documents, `gno_graph_path` for "how are X and Y connected?", `gno_links`/`gno_backlinks` for one-document link expansion, and `gno_similar` for semantic neighbors. Prefer explicit or typed edges over inferred, ambiguous, or similarity edges when confidence matters.
+5. Use graph/link expansion for relationship context: `gno_graph_query` for typed relationship traversal, `gno_graph_neighbors` for nearby documents, `gno_graph_path` for "how are X and Y connected?", `gno_links`/`gno_backlinks` for one-document link expansion, and `gno_similar` for semantic neighbors. Prefer explicit or typed edges over inferred, ambiguous, or similarity edges when confidence matters. Plain `[[Note]]` links resolve across collections of the same vault (link workspace), so backlinks and impact can name other collections; pass `collection`/`collections` to keep results inside the user's scope. Never tell users to rewrite links into `[[collection:Note]]` to make them resolve.
 6. Use `gno_query_diagnose` when a known target document should have appeared but did not; it reports BM25/vector/fusion/graph/rerank stage presence and filter state.
 7. Use `gno_get` with `fromLine`/`lineCount` for targeted reads, or `gno_multi_get` to batch top refs.
 8. Use `gno_section` only when you need a durable section locator or must re-resolve one after edits. Prefer search → `gno_get` for ordinary retrieval. `action=create` needs `ref` plus exactly one of `anchor`|`line`; `action=resolve` needs `ref` plus `target`. Cite or open content only for `exact`/`recovered` results, then follow the tool's ready-to-use `gno_get` guidance (`fromLine = lineStart`; `lineCount = lineEnd - lineStart + 1`). Never navigate or cite `ambiguous`/`stale`/`missing`.
@@ -418,6 +427,7 @@ changed source:
 gno changes --since 2026-07-20T00:00:00Z --json
 gno diff gno://notes/plan.md --json
 gno impact gno://notes/plan.md --max-depth 3 --json
+gno impact gno://notes/plan.md --collection notes --json   # stay in scope
 ```
 
 Treat cursors and change IDs as opaque. Journal results are bounded,

package/assets/skill/cli-reference.md

@@ -594,7 +594,8 @@ gno impact gno://notes/plan.md --max-depth 3 --json
 `changes` accepts an ISO time or opaque cursor and optional collection/limit.
 `diff` reports structural headings, links, and typed-relationship changes for
 one retained change. `impact` follows inbound evidence edges with explicit
-depth/node/edge/frontier/visited bounds. Expired journal history is reported,
+depth/node/edge/frontier/visited bounds; repeat `-c, --collection <name>` to
+keep the traversal (and every path step) inside those collections. Expired journal history is reported,
 not reconstructed.
 
 ## Private Retrieval Traces
@@ -679,6 +680,12 @@ gno backlinks <ref> [options]
 `--relation <type>` to query semantic typed edges instead of positional
 wiki/markdown links. Do not combine `--type` with `--edge-type`.
 
+Plain `[[Note]]` and `[[Folder/Note]]` links resolve across all collections
+that share one vault (the nearest `.obsidian` folder, or `workspaceRoot` in
+the config): exact path, then same folder, then shallowest file; ties stay
+unresolved and are listed by `gno audit links`. JSON names the other
+collection (`resolvedCollection` for links, `sourceCollection` for backlinks).
+
 ### gno graph query
 
 Bounded typed-edge traversal from a document.

package/assets/skill/examples.md

@@ -314,7 +314,8 @@ gno similar gno://notes/auth.md --cross-collection
 # In your documents:
 
 See [[API Design]] for details.
-Check [[work:Project Plan]] for cross-collection link.
+Check [[Projects/Plan]] for a path link (resolves across collections of one vault).
+Check [[work:Project Plan]] to name the target collection explicitly.
 Read [[Security#OAuth]] for specific section.
 ```
 

package/assets/skill/mcp-reference.md

@@ -144,7 +144,9 @@ lives in `structuredContent`; its text projection is deliberately compact and
 should not be expanded back into duplicate model context.
 
 Use `gno_changes`, `gno_diff`, and `gno_impact` for retained metadata history
-and bounded dependency questions. Use `gno_trace_list` and `gno_trace_show` for
+and bounded dependency questions. `gno_impact` takes optional `collections`
+to keep the traversal in scope; links in a vault can resolve across
+collections, so unscoped graph tools cover every collection. Use `gno_trace_list` and `gno_trace_show` for
 private local diagnostics. Invoke `gno_trace_label` only when the user
 explicitly provides a relevant, irrelevant, or missing-expected judgment.
 Trace export/replay/delete/purge and saved-Capsule watch lifecycle remain

package/assets/skill/recipes/memory-scoped-recall.md

@@ -32,11 +32,16 @@ the flag.
      Superseded records never appear.
    - `budget.omitted` > 0: facts matched but did not fit. Narrow the query
      or raise the budget.
-   - Empty `facts` plus a `hint`: nothing is stored in scope yet. Say so;
-     the hint names `gno remember` for when the user wants to store one.
-   - `retrieval.mode: "lexical"`: the collection has no cached embeddings, so
-     the query matched every term. Rephrase to the fact's own words, or embed
-     the collection (`gno embed <collection>`) for question-shaped queries.
+   - Empty `facts` plus a `hint`: the hint says why. "No memories in scope
+     yet" means nothing is stored there; "No memories in scope matched this
+     query" means facts exist but none shares a word with the question, so
+     retry with other words before concluding memory is silent. The hints
+     name `gno remember` for when the user wants to store one.
+   - `retrieval.mode: "lexical"`: the collection has no cached embeddings.
+     Questions still work (question words are dropped and facts sharing a
+     content word are returned), but a paraphrase with no word in common
+     misses; embed the collection (`gno embed <collection>`) to match by
+     meaning.
 
 3. Answer from the facts and cite each by its `gno://` URI. When memory is
    silent or stale, fall through to `gno search` / `gno query` on the

package/browser-extension/artifacts/gno-browser-clipper-v2.8.1.zip.sha256

@@ -0,0 +1 @@
+3f23832d1ed8b19e24eef29d6ea888a8c84009e39c61981d0b665c95cd7d792e  gno-browser-clipper-v2.8.1.zip

package/browser-extension/dist/manifest.json

@@ -21,5 +21,5 @@
   "content_security_policy": {
     "extension_pages": "script-src 'self'; object-src 'none'; connect-src http://127.0.0.1:*"
   },
-  "version": "2.7.1"
+  "version": "2.8.1"
 }

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@gmickel/gno",
-  "version": "2.7.1",
+  "version": "2.8.1",
   "description": "Local semantic search for your documents. Index Markdown, PDF, and Office files with hybrid BM25 + vector search.",
   "keywords": [
     "embeddings",

package/spec/cli.md

@@ -240,7 +240,12 @@ the model with `id`, `model`, `dimensions`, `state` (`active`|`shadow`),
 `provenance` (building runtime, e.g. `CUDA, Bun 1.4.2`),
 `compatibleRuntimes` (runtimes that read it) and `incompatibleRuntimes`.
 Terminal output prints a `Vector partitions:` block
-unless there is exactly one healthy partition. Per-collection chunk totals remain deduplicated by canonical chunk;
+unless there is exactly one healthy partition. `collections`, `totalDocuments`
+and `totalChunks` cover configured collections only: a collection removed from
+config (for example by `gno collection remove`) is not reported even while its
+rows remain in the index until the next `gno update`. `totalChunks` counts the
+distinct chunks of active documents; chunks left by deleted documents do not
+count. Per-collection chunk totals remain deduplicated by canonical chunk;
 embedded counts require matching current inputs for every active owner within
 that collection. Status reads persisted identity and coverage without loading
 models. Legacy storage remains the fallback before variant authority; ambiguous
@@ -263,6 +268,8 @@ gno status [--json|--md]
     {
       "name": "work",
       "path": "/path",
+      "workspaceRoot": "/vault",
+      "workspaceSource": "detected",
       "documentCount": 100,
       "chunkCount": 500,
       "embeddedCount": 500
@@ -624,9 +631,11 @@ refuses descent into dataless or availability-unknown directories
 (`DATALESS_DIRECTORY` skip, or the fail-closed codes above) and preserves
 previously indexed descendants under those unproven prefixes rather than
 marking them inactive. Eligible files have no availability `errorCode`.
-Evidence-qualified scope: Google Drive, iCloud Drive, and OneDrive only for the
-tested OS/provider configuration; OneDrive only for both validated immediate
-SharePoint library roots. No Windows/Linux guarantee; metadata/provider
+Evidence-qualified scope: Google Drive (`My Drive` and immediate Shared drives
+under `Shared drives/<drive>`; the `Shared drives` folder itself is
+unsupported), iCloud Drive, and OneDrive only for the tested OS/provider
+configuration; OneDrive only for the two validated immediate SharePoint library
+roots. No Windows/Linux guarantee; metadata/provider
 bookkeeping may occur; GNO does not pin/evict/download as product behavior.
 Source availability is distinct from `egressPolicy` (where derived content may
 travel).
@@ -886,6 +895,15 @@ projection.
 content-free decision, lineage, partial disclosure, audit metadata, and
 remediation contract.
 
+`workspaceSource` is `detected` (nearest `.obsidian/` ancestor-or-self of the
+collection root), `configured` (`workspaceRoot` setting), `disabled`
+(`workspaceRoot: false`), `unavailable` (the root or an ancestor could not be
+inspected; links stay collection-scoped) or `none`. `workspaceRoot` is the
+effective workspace root, present only for `detected`/`configured`, and is a
+same-host-only field like `path`. Terminal output adds a `Link workspace:` line
+per workspace collection. `gno collection list` reports the same per
+collection as `effectiveWorkspaceRoot` and `workspaceSource`.
+
 ### gno audit
 
 Read-only, offline knowledge-integrity audits. This command is distinct from
@@ -901,8 +919,17 @@ gno audit [links|provenance|freshness|all] [--collection <name>...] \
   [--json] [--output <path>]
 ```
 
-The default category is `all`; `--max-findings` defaults to 100 and is bounded
-to 1–1000. Truncation limits returned findings but preserves exact totals.
+The default category is `all`; `--max-findings` defaults to 100 and accepts an
+integer from 1 to 100000 or `all`. `all` returns every finding of the bounded
+audit snapshot, and per-rule caps follow the same value. Any other value
+(zero, negative, non-integer, above 100000) is a validation error (exit 1)
+that states the accepted range. Truncation limits returned findings but
+preserves exact totals. The report's `truncation` block reports three separate
+conditions: `findingsTruncated` (the finding cap cut the list),
+`snapshotTruncated` (the bounded snapshot of 50,000 documents/links was cut, so
+totals cover the snapshot only and are not complete-index totals), and
+`evidenceTruncated` (evidence items, tied-candidate lists, or evidence detail
+text were shortened).
 `--output` writes only the requested report artifact with local file
 permissions. Human output renders the same report represented by
 `audit-report.schema.json`.
@@ -924,6 +951,41 @@ healthy.
 - `4` — complete report with findings
 - `5` — partial, inconclusive, unavailable, or changed-during-audit evidence
 
+Link findings resolve with the workspace-aware resolver (see
+docs/ARCHITECTURE.md "Resolution"). Their evidence `detail` JSON carries
+separate `referenceKind` (`wiki-name`, `wiki-path`, `explicit-collection`,
+`markdown`), `resolutionStatus` (`unresolved`, `ambiguous`, `outside-index`)
+and `resolvedScope` (`same-collection`, `cross-collection`,
+`explicit-collection`, or null when unresolved). Ambiguous workspace links add `candidateCount` and
+`candidates` (tied candidate URIs in canonical path order); in a
+collection-scoped audit, candidates outside the requested collections are
+counted in `candidatesWithheld` and never named, and a list longer than the
+detail bound sets `candidatesTruncated` (and `truncation.evidenceTruncated`).
+Orphans stay "no incoming or outgoing resolved links", with connectivity drawn
+from the whole index even when the audited documents are scoped; a tied link
+connects nothing.
+
+A plain wiki link or embed from a document in a link workspace whose target
+is not an indexed document but exists as a file inside the workspace root (a
+non-Markdown attachment, a note in an unindexed folder, a note excluded by a
+collection pattern) is not reported by `links.local-targets`. It is an `info`
+finding of `links.outside-index` (evidence kind `outside-index-target`,
+`resolutionStatus: "outside-index"`). That rule stays `pass`, so these
+findings never change the exit code. The check is existence-only: each
+involved workspace root is listed once per run (hidden files and folders
+skipped, symlinked folders not descended, a symlink counted only when it
+resolves to a regular file inside the workspace, at most 200,000 files), no
+file is opened, and no graph edge is created. Targets match files by the workspace
+link rules; a non-Markdown target needs its extension, `.md` is optional. A
+target missing from the listing stays unresolved; an unreadable folder or the
+file bound makes the listing incomplete, which the `links.outside-index` rule
+message states once.
+
+The wiki link parser reads Obsidian's table-escaped alias `[[Note\|Alias]]`
+as target `Note` with alias `Alias`. Markdown link text may contain balanced
+square brackets (`[see [1]](note.md)`); a destination containing a square
+bracket is not a link.
+
 The JSON contract is versioned as `gno://schemas/audit-report@1.0`. Finding IDs
 are stable SHA-256 identities derived from rule, normalized subject/location,
 and load-bearing evidence. Wall-clock timing and traversal order do not affect
@@ -1978,6 +2040,12 @@ gno recall <query> --scope <scope> [--scope <scope>...] [--collection <name>] [-
   `retrieval.semanticUnavailable` explaining why). Query expansion, graph
   expansion, and reranking are disabled. Scope and supersession filtering run
   inside the retrieval query; superseded facts are never returned.
+- The lexical leg searches the query's content words (question and function
+  words dropped; quoted phrases and `-term` exclusions kept). Facts containing
+  every content word are returned first; when none does, facts sharing any
+  content word are returned in BM25 order, dropping those scoring below 0.1
+  of the best match.
+- The budget is filled in retrieval-rank order.
 - Budget: at most `--max-facts` facts (default 8) under `--max-tokens`
   (default 512). Both must be positive integers. Recall never downloads a
   model.
@@ -1987,8 +2055,11 @@ gno recall <query> --scope <scope> [--scope <scope>...] [--collection <name>] [-
   (`caller`, `session`, `issuedAt`, `memoryIds`, `spanHashes`, `digest`) plus
   `budget` and `retrieval`. Derived output inherits the strictest source
   egress policy (`egressLineage`).
-- Empty recall prints the self-teaching line naming `gno remember`
-  (`hint` in JSON) and exits 0.
+- Empty recall prints a hint (`hint` in JSON) and exits 0. The hint
+  distinguishes an empty scope (`No memories in scope yet. Store one with:
+gno remember ...`), a populated scope with no match (`No memories in scope
+matched this query. ...`, also naming `gno remember`), and matches that did
+  not fit the token budget.
 
 **Output:** `--json` prints the shared `RecallResult`. Terminal output lists
 numbered facts with URI, text, scopes, hash, and identity, then `Budget:`,
@@ -4179,13 +4250,19 @@ Find active documents that depend on one document through inbound typed,
 wiki-link, or Markdown-link edges.
 
 ```bash
-gno impact <doc> [--max-depth <n>] [--max-nodes <n>] [--max-edges <n>] [--frontier-limit <n>] [--visited-limit <n>] [--json]
+gno impact <doc> [-c, --collection <name>...] [--max-depth <n>] [--max-nodes <n>] [--max-edges <n>] [--frontier-limit <n>] [--visited-limit <n>] [--json]
 ```
 
 The traversal is cycle-safe and enforces depth, node, edge, frontier, and
 visited-row caps. Every impacted document includes one deterministic
 dependency-to-root evidence path. JSON output uses `impact.schema.json`.
 
+`--collection` (repeatable) limits the traversal to those collections: only
+their documents are visited, returned, or used as a path step, so a document
+outside the scope never bridges two in-scope documents. Omitted means every
+indexed collection. An unknown collection, or a `<doc>` outside the requested
+collections, is a validation error (exit 1).
+
 **Exit Codes:**
 
 - 0: Success

package/spec/db/schema.sql

@@ -172,7 +172,6 @@ CREATE TABLE IF NOT EXISTS documents (
 );
 
 CREATE INDEX IF NOT EXISTS idx_documents_collection ON documents(collection);
-CREATE INDEX IF NOT EXISTS idx_documents_active ON documents(active);
 CREATE INDEX IF NOT EXISTS idx_documents_mirror_hash ON documents(mirror_hash);
 CREATE INDEX IF NOT EXISTS idx_documents_docid ON documents(docid);
 CREATE INDEX IF NOT EXISTS idx_documents_uri ON documents(uri);

package/spec/mcp.md

@@ -351,7 +351,8 @@ Collection names are case-insensitive on input and normalized to lowercase in re
 `gno_audit` is a read-only, offline tool that returns the same
 `gno://schemas/audit-report@1.0` report as `gno audit`. Its closed input accepts
 `category` (`links`, `provenance`, `freshness`, or `all`), collection/path/tag
-filters, and `maxFindings` (1–1000). `maxAgeDays`, `orphanRoots`, and
+filters, and `maxFindings` (an integer from 1 to 100000, or `"all"` for every
+finding of the bounded snapshot; default 100). `maxAgeDays`, `orphanRoots`, and
 `orphanIgnorePrefixes` are optional explicit policy inputs. Request
 cancellation returns a partial report rather than a false clean result. The
 tool is annotated with
@@ -360,7 +361,10 @@ modifies source files, config, index rows, graph edges, daemon state, or a
 persisted audit baseline.
 
 Rules report `pass`, `fail`, `skip`, `unavailable`, or `inconclusive`; reports
-are `complete`, `partial`, `changed_during_audit`, or `failed`. Unavailable or
+are `complete`, `partial`, `changed_during_audit`, or `failed`. A passing rule
+may carry `info` findings (`links.outside-index`: link targets that exist as
+files in the link workspace but are not indexed); only `warning` and `error`
+findings fail a rule. Link semantics match `gno audit` (spec/cli.md). Unavailable or
 changing evidence never appears healthy. Stable finding IDs are derived from
 rule, normalized subject/location, and evidence fingerprint. Responses are
 canonically ordered and bounded while retaining exact pre-truncation totals.
@@ -1458,8 +1462,9 @@ Budgeted, cited, current-state recall from a memory-managed collection
   the vector leg did not run
 - `egressLineage` — strictest source policy across returned facts (absent when
   empty)
-- `hint` — self-teaching line naming `gno remember`, present only when no fact
-  was returned
+- `hint` — present only when no fact was returned; says why: the scope holds
+  no current fact (names `gno remember`), nothing in scope matched the query,
+  or the matches did not fit `maxTokens`
 
 **Identity:** `caller` is the MCP client implementation name from the
 `initialize` handshake (`mcp` when absent); `session` is the Streamable HTTP
@@ -1841,8 +1846,15 @@ pending), `sourceUnavailable`, `staleParser` and last import time, and the
 switches, pending and running work, last run, last success, next due time,
 recovery action). The result contains no host paths.
 
+Each call (and each `gno_sessions_import` call) first re-reads the server's
+config file and adopts it when it changed, so source changes made by another
+process show without a restart. An unchanged file is a no-op; a source-only
+change keeps open HTTP sessions.
+
 **Errors:** `SESSIONS_NOT_CONFIGURED` when the server's config has no
-`sessions` block.
+`sessions` block; `SESSIONS_RUNTIME_FAILURE` when the config file cannot be
+read (never answered from the stale config); `SESSIONS_BINDING_MISMATCH` when
+the file is now bound to a different index (not adopted).
 
 ### gno_sessions_import
 
@@ -2177,7 +2189,7 @@ Find semantically similar documents using vector embeddings.
 **Algorithm:**
 
 1. Get all chunks for the source document
-2. Retrieve embeddings for each chunk from content_vectors
+2. Retrieve each chunk's stored embedding from the active vector partition (legacy `content_vectors` only before any partition activates); no model is loaded
 3. Compute average embedding across all chunks
 4. Search for nearest neighbors using sqlite-vec
 5. Exclude self and filter by collection if not crossCollection
@@ -2371,7 +2383,14 @@ structure is derived from `structureDelta.truncated`.
 
 ### gno_impact
 
-Read-only inbound dependency traversal for `ref`. Inputs `maxDepth`,
+Read-only inbound dependency traversal for `ref`. Optional `collections`
+(array of collection names) limits the traversal as `gno impact --collection`
+does; omitted means every collection. Over Streamable HTTP, egress policy is
+checked on `collections` (every collection when omitted, because the result
+can reach any collection a link resolves into) and on the collection of `ref`;
+a docid `ref` is checked against every collection. The same rule applies to
+`gno_backlinks` and `gno_graph*` refs, and `gno_similar` with
+`crossCollection: true` is checked against every collection. Inputs `maxDepth`,
 `maxNodes`, `maxEdges`, `frontierLimit`, and `visitedLimit` use the same bounds
 as CLI/REST/SDK. Structured content is `impact.schema.json`; each impacted
 document includes a deterministic evidence path over typed or backlink

package/spec/output-schemas/audit-report.schema.json

@@ -260,13 +260,27 @@
     "truncation": {
       "type": "object",
       "additionalProperties": false,
-      "required": ["findingsTruncated", "maxFindings"],
+      "required": [
+        "findingsTruncated",
+        "maxFindings",
+        "snapshotTruncated",
+        "evidenceTruncated"
+      ],
       "properties": {
         "findingsTruncated": { "type": "boolean" },
         "maxFindings": {
-          "type": "integer",
-          "minimum": 1,
-          "maximum": 1000
+          "oneOf": [
+            { "type": "integer", "minimum": 1, "maximum": 100000 },
+            { "const": "all" }
+          ]
+        },
+        "snapshotTruncated": {
+          "type": "boolean",
+          "description": "The bounded audit snapshot was cut; finding totals cover the snapshot only, not the whole index"
+        },
+        "evidenceTruncated": {
+          "type": "boolean",
+          "description": "Evidence items, candidate lists, or evidence detail text were shortened"
         }
       }
     },

package/spec/output-schemas/backlinks.schema.json

@@ -28,6 +28,10 @@
                 "type": "string",
                 "description": "Source document title"
               },
+              "sourceCollection": {
+                "type": "string",
+                "description": "Collection of the linking document (may differ from the target's in a link workspace)"
+              },
               "linkText": {
                 "type": "string",
                 "description": "Link display text"

package/spec/output-schemas/collection-list.schema.json

@@ -39,6 +39,19 @@
       "languageHint": {
         "type": "string",
         "description": "BCP-47 language hint for the collection"
+      },
+      "workspaceRoot": {
+        "description": "Configured link workspace: an absolute path joins the collection to that workspace root; false keeps links inside the collection. Absent means auto-detect. Same-host callers only: omitted for a remote REST caller.",
+        "oneOf": [{ "type": "string" }, { "const": false }]
+      },
+      "effectiveWorkspaceRoot": {
+        "type": "string",
+        "description": "CLI list only: effective link workspace root (nearest .obsidian folder, or the configured root). Absent when links stay inside the collection."
+      },
+      "workspaceSource": {
+        "type": "string",
+        "enum": ["none", "detected", "configured", "disabled", "unavailable"],
+        "description": "CLI list only: how the link workspace was set"
       }
     }
   }

package/spec/output-schemas/graph.schema.json

@@ -92,6 +92,8 @@
                 "enum": [
                   "exact-title",
                   "exact-path",
+                  "exact-name",
+                  "tie-break",
                   "path-fallback",
                   "ambiguous-fallback",
                   "similarity"

package/spec/output-schemas/links-list.schema.json

@@ -78,6 +78,10 @@
                 "type": "string",
                 "description": "URI of resolved target document"
               },
+              "resolvedCollection": {
+                "type": "string",
+                "description": "Collection of the resolved target (may differ from the source's in a link workspace)"
+              },
               "resolvedTitle": {
                 "type": "string",
                 "description": "Title of resolved target document"

package/spec/output-schemas/memory-recall.schema.json

@@ -38,7 +38,7 @@
     },
     "hint": {
       "type": "string",
-      "description": "Self-teaching line naming `gno remember`; present only when no fact was returned."
+      "description": "Why nothing was returned (empty scope, no match in scope, or matches over the token budget); present only when no fact was returned."
     }
   },
   "$defs": {

package/spec/output-schemas/status.schema.json

@@ -98,7 +98,7 @@
     },
     "collections": {
       "type": "array",
-      "description": "Collection statistics",
+      "description": "Statistics for configured collections; a collection removed from config is omitted even before its index rows are pruned",
       "items": {
         "type": "object",
         "required": ["name", "documentCount", "chunkCount", "embeddedCount"],
@@ -111,6 +111,21 @@
             "type": "string",
             "description": "Collection root path. Same-host callers only: omitted for a remote REST or HTTP MCP caller."
           },
+          "workspaceRoot": {
+            "type": "string",
+            "description": "Effective link workspace root: plain wiki links resolve across every collection under it. Present only when the collection is in a workspace. Same-host callers only: omitted for a remote REST or HTTP MCP caller."
+          },
+          "workspaceSource": {
+            "type": "string",
+            "enum": [
+              "none",
+              "detected",
+              "configured",
+              "disabled",
+              "unavailable"
+            ],
+            "description": "How the link workspace was set: detected (nearest .obsidian folder), configured (workspaceRoot setting), disabled (workspaceRoot: false), unavailable (root could not be inspected; links stay collection-scoped), none (no workspace)"
+          },
           "documentCount": {
             "type": "integer",
             "description": "Number of active documents",
@@ -131,12 +146,12 @@
     },
     "totalDocuments": {
       "type": "integer",
-      "description": "Total documents across all collections",
+      "description": "Active documents across configured collections",
       "minimum": 0
     },
     "totalChunks": {
       "type": "integer",
-      "description": "Total chunks across all collections",
+      "description": "Distinct chunks of active documents across configured collections",
       "minimum": 0
     },
     "embeddingBacklog": {

package/src/cli/commands/audit.ts

@@ -14,8 +14,11 @@ import type { WorkspaceAuditProgress } from "../../core/audit-workspace";
 import { getIndexDbPath } from "../../app/constants";
 import { loadConfig } from "../../config";
 import {
+  AUDIT_MAX_FINDINGS_RANGE_MESSAGE,
   auditExitCode,
   AUDIT_CATEGORIES,
+  parseAuditMaxFindingsInput,
+  resolveAuditMaxFindings,
   serializeAuditReportCanonical,
 } from "../../core/audit";
 import { runWorkspaceAudit } from "../../core/audit-workspace";
@@ -31,7 +34,8 @@ export interface AuditCommandOptions {
   collections?: string[];
   paths?: string[];
   tags?: string[];
-  maxFindings?: number;
+  /** Positive integer, `all`, or the raw CLI string for either. */
+  maxFindings?: number | string;
   maxAgeDays?: number;
   orphanRoots?: string[];
   orphanIgnorePrefixes?: string[];
@@ -80,11 +84,18 @@ export const audit = async (
       error: "category must be links, provenance, freshness, or all",
     };
   }
-  if (invalidPositiveInteger(options.maxFindings)) {
+  const maxFindings =
+    typeof options.maxFindings === "string"
+      ? parseAuditMaxFindingsInput(options.maxFindings)
+      : options.maxFindings;
+  if (
+    (options.maxFindings !== undefined && maxFindings === undefined) ||
+    !resolveAuditMaxFindings(maxFindings).ok
+  ) {
     return {
       success: false,
       invalid: true,
-      error: "maxFindings must be a positive integer",
+      error: AUDIT_MAX_FINDINGS_RANGE_MESSAGE,
     };
   }
   if (invalidPositiveInteger(options.maxAgeDays)) {
@@ -155,7 +166,7 @@ export const audit = async (
       collectionFilters: requestedCollections,
       pathFilters: options.paths,
       tagFilters: requestedTags,
-      maxFindings: options.maxFindings,
+      maxFindings,
       agePolicy:
         options.maxAgeDays === undefined
           ? undefined
@@ -188,6 +199,14 @@ export const formatAuditReport = (
     `Categories: ${report.scope.categories.join(", ")}`,
     `Rules: ${report.counts.rules.total} (${report.counts.rules.fail} failed, ${report.counts.rules.unavailable} unavailable, ${report.counts.rules.inconclusive} inconclusive)`,
     `Findings: ${report.counts.findings.total}${report.counts.findings.truncated ? ` (${report.counts.findings.returned} shown)` : ""}`,
+    ...(report.truncation.snapshotTruncated
+      ? [
+          "Snapshot: truncated; totals cover the bounded audit snapshot, not the whole index",
+        ]
+      : []),
+    ...(report.truncation.evidenceTruncated
+      ? ["Evidence: some finding evidence was shortened"]
+      : []),
     `Examined: ${report.counts.examined.documents} document/rule observations`,
     `Duration: ${report.durationMs}ms`,
   ];