npm - @cerefox/memory - Versions diffs - 1.6.0 → 1.7.0 - Mend

@cerefox/memory 1.6.0 → 1.7.0

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.

Files changed (27) hide show

package/AGENT_GUIDE.md +79 -5
package/AGENT_QUICK_REFERENCE.md +9 -7
package/README.md +5 -3
package/dist/bin/cerefox.js +591 -227
package/dist/frontend/assets/index-D8E0mTnp.js +121 -0
package/dist/frontend/assets/index-D8E0mTnp.js.map +1 -0
package/dist/frontend/assets/{index-DiDeaiM6.css → index-Dm_zCch4.css} +1 -1
package/dist/frontend/index.html +2 -2
package/dist/server-assets/_shared/ef-meta/index.ts +3 -3
package/dist/server-assets/_shared/mcp-tools/_utils.ts +41 -0
package/dist/server-assets/_shared/mcp-tools/delete-document.ts +181 -0
package/dist/server-assets/_shared/mcp-tools/get-help-content.ts +4 -4
package/dist/server-assets/_shared/mcp-tools/index.ts +8 -1
package/dist/server-assets/_shared/mcp-tools/ingest.ts +75 -8
package/dist/server-assets/_shared/mcp-tools/partial-edits.ts +43 -4
package/dist/server-assets/_shared/mcp-tools/restore-document.ts +130 -0
package/dist/server-assets/db/migrations/0024_mcp_delete_document.sql +26 -0
package/dist/server-assets/db/rpcs.sql +207 -27
package/dist/server-assets/db/schema.sql +1 -1
package/dist/server-assets/supabase/functions/cerefox-ingest/index.ts +26 -0
package/docs/guides/access-paths.md +24 -15
package/docs/guides/cli.md +13 -7
package/docs/guides/connect-agents.md +41 -17
package/docs/guides/operational-cost.md +1 -1
package/package.json +1 -1
package/dist/frontend/assets/index-DcWOeGAh.js +0 -121
package/dist/frontend/assets/index-DcWOeGAh.js.map +0 -1

package/dist/frontend/index.html CHANGED Viewed

@@ -15,8 +15,8 @@
       href="https://fonts.googleapis.com/css2?family=Geist:wght@300;400;500;600;700&display=swap"
     />
     <title>Cerefox</title>
-    <script type="module" crossorigin src="/app/assets/index-DcWOeGAh.js"></script>
-    <link rel="stylesheet" crossorigin href="/app/assets/index-DiDeaiM6.css">
+    <script type="module" crossorigin src="/app/assets/index-D8E0mTnp.js"></script>
+    <link rel="stylesheet" crossorigin href="/app/assets/index-Dm_zCch4.css">
   </head>
   <body>
     <div id="root"></div>

package/dist/server-assets/_shared/ef-meta/index.ts CHANGED Viewed

@@ -18,7 +18,7 @@
  * doesn't touch `supabase/functions/` leaves it alone).
  */
-export const EF_VERSION = "1.6.0";
+export const EF_VERSION = "1.7.0";
 /**
  * The Cerefox RELEASE version — what `cerefox --version` reports and what npm
@@ -36,7 +36,7 @@ export const EF_VERSION = "1.6.0";
  * is imported by the Deno Edge Functions, which cannot reach into the npm
  * package.
  */
-export const CEREFOX_VERSION = "1.6.0";
+export const CEREFOX_VERSION = "1.7.0";
 /**
  * The most recent version whose EF-side SOURCE actually changed (#127).
@@ -46,7 +46,7 @@ export const CEREFOX_VERSION = "1.6.0";
  * `cut_release.ts` ONLY when EF source changed since the last tag; doctor
  * uses it to stay silent on label-only drift.
  */
-export const EF_LAST_CHANGED = "1.6.0";
+export const EF_LAST_CHANGED = "1.7.0";
 /**
  * The 8 peer EFs the cerefox-mcp aggregator probes (excludes cerefox-mcp

package/dist/server-assets/_shared/mcp-tools/_utils.ts CHANGED Viewed

@@ -181,6 +181,47 @@ export interface LogUsageParams {
  *  Differs from the EF's `logUsage` only in that `accessPath` is a required
  *  parameter (was hardcoded to `"remote-mcp"` in the EF) so the local TS
  *  MCP server can pass `"local-mcp"` for the same call site. */
+/** Pull the two hashes out of an RPC `CEREFOX_CONFLICT` message.
+ *
+ *  The ONE site coupled to the SQL `RAISE` wording ("expected hash %, current
+ *  hash %"). Three tools (ingest, edit, delete) rephrase conflicts for agents;
+ *  before this helper each carried its own copy of these regexes, and a
+ *  wording change in rpcs.sql would have had to be mirrored three times, with
+ *  a missed one silently degrading that tool's conflict output to "unknown". */
+export function extractConflictHashes(message: string): { expected: string; current: string } {
+  return {
+    expected: message.match(/expected hash ([0-9a-f]{64})/)?.[1] ?? "unknown",
+    current: message.match(/current hash ([0-9a-f]{64})/)?.[1] ?? "unknown",
+  };
+}
+/** Does this RPC error mean the function (or this signature of it) is not on
+ *  the server — i.e. the client is newer than the deployed schema?
+ *
+ *  The ONE site coupled to how that condition surfaces: PostgREST's PGRST202
+ *  ("Could not find the function … in the schema cache") and Postgres' 42883
+ *  ("function … does not exist", seen through connections that bypass
+ *  PostgREST or during a deploy's DROP/CREATE window). Five call sites
+ *  (delete/restore MCP handlers, delete/restore CLI verbs, partial edits)
+ *  each carried a hand-rolled subset of these predicates before this helper. */
+/** Is this RPC error the delete/restore RPCs' "Document % not found"?
+ *
+ *  Anchored on the SQLSTATE (22023, invalid_parameter_value — PostgREST
+ *  passes it through as `error.code`) AND the prose, because 22023 alone is
+ *  shared with other validation raises (zero chunks, token required) and the
+ *  prose alone would match any gateway error containing "not found". One
+ *  site instead of four hand-rolled substring checks. */
+export function isDocumentNotFoundError(error: { code?: string; message?: string }): boolean {
+  return error.code === "22023" && /not found/i.test(error.message ?? "");
+}
+export function isMissingFunctionError(message: string, fnName: string): boolean {
+  return (
+    (message.includes("Could not find the function") && message.includes(fnName)) ||
+    (message.includes("does not exist") && message.includes(fnName))
+  );
+}
 export function logUsage(supabase: MCPSupabaseClient, params: LogUsageParams): void {
   Promise.resolve(
     supabase.rpc("cerefox_log_usage", {

package/dist/server-assets/_shared/mcp-tools/delete-document.ts ADDED Viewed

@@ -0,0 +1,181 @@
+/**
+ * `cerefox_delete_document` — soft-delete a document over MCP (#208).
+ *
+ * Closes a parity gap, not a policy: the trust model (access-paths.md →
+ * "Destructive operations and the trust model") always sanctioned agent
+ * soft-delete — audited, recoverable from the web-UI trash — but the tool was
+ * never built, so agents had no path where the CLI had `document delete`.
+ *
+ * The MCP surface REQUIRES `expected_content_hash` where the CLI requires an
+ * interactive y/N: both are proof-of-intent, and an agent's proof is that it
+ * READ the thing it is deleting. The CAS itself lives in the RPC under the
+ * same FOR UPDATE lock as the ingest CAS — only the "required" is enforced
+ * here, because it is an MCP-surface contract, not a data-integrity one
+ * (the CLI legitimately deletes without a hash).
+ *
+ * Its inverse is `cerefox_restore_document` (#210, same release). What has no
+ * agent surface is **permanent purge** — the one action that destroys data
+ * keeps its web-UI human-in-the-loop confirmation.
+ */
+import type { MCPSupabaseClient } from "./types.ts";
+import {
+  extractConflictHashes,
+  isDocumentNotFoundError,
+  isMissingFunctionError,
+  logUsage,
+} from "./_utils.ts";
+import { McpInvalidParams, type ToolContext, type ToolDefinition } from "./types.ts";
+/** Agent-first instructions for a stale-hash conflict on delete. */
+function conflictError(documentId: string, expectedHash: string, currentHash: string): Error {
+  return new Error(
+    `Conflict: document ${documentId} changed since you read it ` +
+      `(your base hash: ${expectedHash}, current hash: ${currentHash}). ` +
+      `Someone wrote to this document after you decided to delete it. ` +
+      `To resolve: (1) cerefox_get_document("${documentId}") to see the current ` +
+      `content, (2) check the deletion is still warranted — the new content may ` +
+      `change your mind, (3) if it is, retry with expected_content_hash set to ` +
+      `the new hash. Do not delete blindly — the change may be another writer's work.`,
+  );
+}
+async function handler(
+  supabase: MCPSupabaseClient,
+  args: Record<string, unknown>,
+  ctx: ToolContext,
+): Promise<string> {
+  const document_id = args.document_id as string | undefined;
+  // Trimmed: a correct hash with a stray newline is not a stale one. The RPC
+  // trims too; doing it here keeps the fast-fail message honest as well.
+  const expected_content_hash = (args.expected_content_hash as string | undefined)?.trim();
+  const reason = args.reason as string | undefined;
+  if (!document_id) throw new McpInvalidParams("document_id is required");
+  if (!expected_content_hash) {
+    // The MCP analogue of the CLI's confirmation prompt: prove you read what
+    // you are deleting. There is no last_write_wins here on purpose — a
+    // delete with no evidence of a read has no legitimate agent use case.
+    throw new McpInvalidParams(
+      "expected_content_hash is required: the content_hash of the document as you read it " +
+        "(returned by cerefox_get_document, cerefox_search, and cerefox_metadata_search). " +
+        "If you have not read the document, read it first — deletion requires knowing what " +
+        "you are deleting.",
+    );
+  }
+  const author = (args.author as string | undefined) ?? (args.requestor as string | undefined);
+  // Derived from the transport, never taken from the caller: an agent must not
+  // be able to record itself as a user. Matches the partial-edit handlers.
+  const authorType = ctx.accessPath === "cli" ? "user" : "agent";
+  const { data, error } = await supabase.rpc("cerefox_delete_document", {
+    p_document_id: document_id,
+    p_author: author ?? "unknown",
+    p_author_type: authorType,
+    p_expected_content_hash: expected_content_hash,
+    p_reason: reason ?? null,
+  });
+  if (error) {
+    const message = error.message ?? String(error);
+    if (message.includes("CEREFOX_CONFLICT")) {
+      const { expected, current } = extractConflictHashes(message);
+      throw conflictError(document_id, expected === "unknown" ? expected_content_hash : expected, current);
+    }
+    if (isMissingFunctionError(message, "cerefox_delete_document")) {
+      throw new Error(
+        `This server is behind: cerefox_delete_document needs schema 0.12.0 or newer. ` +
+          `Run \`cerefox server deploy\`, then retry. (${message})`,
+      );
+    }
+    if (isDocumentNotFoundError(error)) {
+      throw new McpInvalidParams(`Document ${document_id} not found.`);
+    }
+    throw new Error(`RPC error: ${message}`);
+  }
+  const row = data as
+    | {
+        document_id?: string;
+        title?: string;
+        total_chars?: number;
+        deleted_at?: string;
+        already_deleted?: boolean;
+      }
+    | undefined;
+  if (!row) throw new Error("cerefox_delete_document returned no data");
+  if (row.already_deleted) {
+    // Report what happened, not what was asked for: nothing was changed, the
+    // original deletion time stands, and no new audit entry was written. No
+    // usage-log entry either — the RPC's no-op design exists so counters and
+    // the audit log cannot disagree, and logging "delete" here would recreate
+    // exactly that disagreement in the analytics.
+    return (
+      `Document ${document_id} ("${row.title ?? "untitled"}") was ALREADY soft-deleted ` +
+      `at ${row.deleted_at}. No change was made and no audit entry was written.`
+    );
+  }
+  logUsage(supabase, {
+    operation: "delete",
+    accessPath: ctx.accessPath,
+    requestor: args.requestor as string | undefined,
+    document_id,
+    result_count: 1,
+  });
+  return (
+    `Soft-deleted "${row.title ?? "untitled"}" (id: ${document_id}, ` +
+    `${row.total_chars ?? "?"} chars) at ${row.deleted_at}.\n` +
+    `The document is excluded from search but fully recoverable: it sits in the ` +
+    `trash until restored or purged. cerefox_restore_document undoes this if it ` +
+    `was a mistake; permanent purge is human-only (web UI).\n` +
+    `Tell your user what you deleted and why, so they can review it.`
+  );
+}
+export const deleteDocumentTool: ToolDefinition = {
+  name: "cerefox_delete_document",
+  description:
+    "SOFT-delete a document: it leaves search results and lands in the trash, recoverable until a human purges it. Requires expected_content_hash — the content_hash of the document AS YOU READ IT — so a delete always follows a read; if the document changed in between, the call fails with a conflict and you should re-read before deciding again. A mistaken delete can be undone with cerefox_restore_document; permanent purge is human-only (web UI). ALWAYS tell your user what you deleted and why. Pass a short reason — it is recorded in the audit log for the human reviewing the trash. Prefer this over ingesting empty/placeholder content when a document should go away.",
+  annotations: {
+    title: "Delete document (soft, recoverable)",
+    readOnlyHint: false,
+    // Static per tool (spec): a tool that can remove content from search is
+    // destructive even though cerefox_restore_document can bring it back —
+    // and once a human purges the trash, the removal is final.
+    destructiveHint: true,
+    // Re-deleting an already-deleted document changes nothing and says so.
+    idempotentHint: true,
+    openWorldHint: false,
+  },
+  inputSchema: {
+    type: "object",
+    required: ["document_id", "expected_content_hash"],
+    properties: {
+      document_id: { type: "string", description: "UUID of the document to soft-delete." },
+      expected_content_hash: {
+        type: "string",
+        description:
+          "The content_hash of the document as you read it (returned by cerefox_get_document, cerefox_search, and cerefox_metadata_search). Required: a delete must follow a read. A stale hash fails with a conflict — re-read, reconsider, retry with the new hash.",
+      },
+      reason: {
+        type: "string",
+        description:
+          "Why this document is being deleted. Recorded in the audit log entry, where it is the main thing the human reviewing the trash has to go on. Short and specific beats long.",
+      },
+      author: {
+        type: "string",
+        description: "Who is making this change. Recorded in the audit log.",
+      },
+      requestor: {
+        type: "string",
+        description: "Name of the agent or user making this request. Recorded in the usage log.",
+      },
+    },
+  },
+  handler,
+};

package/dist/server-assets/_shared/mcp-tools/get-help-content.ts CHANGED Viewed

@@ -11,17 +11,17 @@
  * docs/specs/polish-and-distribution-design.md §10d.
  */
-export const HELP_FULL = "# Cerefox Knowledge Base -- Agent Quick Reference\n\nCerefox is a persistent, shared knowledge base. You have **17 MCP tools** (16 of them have CLI equivalents — `cerefox_get_help` is MCP-only). For the full guide, search Cerefox for \"How AI Agents Use Cerefox\" or call `cerefox_get_help` to retrieve this content over MCP.\n\n## Tools\n\n| Tool | Purpose | Key params |\n|------|---------|------------|\n| `cerefox_search` | Find documents (hybrid FTS + semantic) | `query` (required), `project_name`, `metadata_filter`, `requestor` |\n| `cerefox_ingest` | Save or update a document | `title`, `content` (required), `document_id` (update by ID), `expected_content_hash` (**required on content updates** — see rule 9), `last_write_wins`, `update_if_exists`, `project_name` (single, non-destructive add on update), `project_names` (list, destructive replace on update), `metadata` (omit on update to keep existing tags; `{}` clears), `author` |\n| `cerefox_insert` | **Add** to a document without resending it. Cannot destroy content. | `document_id`, `text`, `position` (`end_of_document`/`end_of_section`/`after_heading`/`before_heading`), `expected_content_hash` (required), `anchor_heading` (unless `end_of_document`), `section_part` |\n| `cerefox_edit` | **Change** parts of a document: 1..n operations applied atomically | `document_id`, `operations` (`insert`/`replace_section`/`delete_section`/`rename_section`), `expected_content_hash` (required) |\n| `cerefox_get_document` | Get full document by ID (header includes `content_hash` — the update token), or with `outline: true` just its heading paths, sizes and hash, or with `section: \"## Heading\"` one section's text | `document_id` (required), `outline`, `section`, `section_part` |\n| `cerefox_list_versions` | Version history of a document | `document_id` (required) |\n| `cerefox_set_relation` ⚑ | Link two documents (`source --rel_type--> target`) | `source_id`, `target_id`, `rel_type` (required), `metadata`, `author` |\n| `cerefox_delete_relation` ⚑ | Remove a relation | `source_id`, `target_id`, `rel_type` |\n| `cerefox_get_relations` ⚑ | All relations touching a document, both directions | `document_id` |\n| `cerefox_get_neighbors` ⚑ | Walk the graph along ONE relation type | `document_id`, `rel_type` (required), `depth`, `from_time`, `to_time`, `limit` |\n| `cerefox_metadata_search` | Find or list docs by metadata, project, or time (no text query) | `metadata_filter`, `project_name` (list a project's docs), `updated_since`, `include_content` — **at least one** of metadata_filter/project_name/updated_since/created_since |\n| `cerefox_list_metadata_keys` | Discover available metadata keys | (none required) |\n| `cerefox_list_projects` | List all projects | (none required) |\n| `cerefox_set_document_metadata` | Change tags WITHOUT resending content. **Merges** by default; a `null` value removes a key | `document_id`, `metadata` (required), `replace` (rare: set exactly this object), `author` |\n| `cerefox_set_document_projects` | Set doc's project memberships to exactly the given list (destructive replace; metadata-only, no content change) | `document_id`, `project_names` (required) |\n| `cerefox_get_audit_log` | Query write operation history | `document_id`, `author`, `operation`, `since` |\n| `cerefox_get_help` | Retrieve Cerefox conventions (this reference) over MCP. **Call this whenever uncertain.** | `topic` (optional, case-insensitive H2 substring match) |\n\n⚑ **Opt-in — usually absent.** The four relation tools are hidden unless the\noperator enables them (`relations_enabled`). **Trust your own tool list**: if\nthey are not in it, the feature is switched off for this deployment. That is\nnormal, not an error, and not something to work around.\n\n## Editing part of a document (prefer this over re-sending)\n\n**Re-sending a whole document to change part of it is the main way agents lose\ndata.** You have to reproduce the untouched remainder verbatim, and any drift\nsilently rewrites content nobody asked you to touch — which the caller cannot\ndiff. Use the partial-edit tools instead:\n\n1. **Learn the anchors** — `cerefox_get_document(document_id, outline: true)`.\n   Returns heading paths, per-section sizes and the `content_hash`, without the\n   body. The paths it returns are exactly what `anchor_heading` accepts.\n2. **Add** → `cerefox_insert`. `end_of_document` is a plain append;\n   `end_of_section` adds inside a named section. It is structurally incapable of\n   removing anything, so \"I meant to append\" cannot become \"I replaced the file\".\n3. **Look before you overwrite** — `cerefox_get_document(document_id,\n   section: \"## Heading\")` returns exactly the text a `replace_section` on that\n   anchor would destroy. The outline gives you a section's *size*, never its\n   *text*, so on a document you did not write yourself this is the difference\n   between a replace and a blind overwrite.\n4. **Change or remove** → `cerefox_edit`. Put changes that belong together in\n   ONE call: they apply atomically, so a table row and the total it feeds cannot\n   end up disagreeing. To change a single line, `replace_section` on its\n   smallest enclosing heading — that is the intended granularity, not a\n   workaround. To fix a stale heading (`## OPEN TODOs (as of ...)`), use\n   `rename_section`: it changes the heading text and leaves the body and\n   position alone.\n5. All of them require `expected_content_hash` and **have no last-write-wins**. A\n   conflict means someone else changed the document; re-read and decide, do not\n   force it.\n\n**A section runs to the next same-or-higher heading, or to the end of the\ndocument.** So `end_of_document` inserts land inside the *last* section, and\nreplacing or deleting that section removes them too. A large shrink in the\nresponse is your warning; `cerefox_list_versions` has the previous content.\n\n**When an anchor is ambiguous the tool refuses and hands you the options** — a\nrepeated heading returns the qualifying paths, and a section with both its own\ncontent and sub-sections returns both `section_part` choices. That is a\nrecoverable answer, not a failure: retry with what it gave you.\n\n## Essential Rules\n\n1. **Search before ingesting** -- check if the document exists first.\n2. **Prefer ID-based updates** -- pass `document_id` from search results for deterministic updates. Falls back to title-matching with `update_if_exists: true`.\n3. **Set `author`/`requestor`** to your name on every call (e.g., \"Claude Code\", \"archiver\"). On MCP, pass as parameters. On CLI, pass `--author`/`--author-type`/`--requestor` flags, or rely on `CEREFOX_AUTHOR_NAME`/`CEREFOX_AUTHOR_TYPE`/`CEREFOX_REQUESTOR_NAME` env vars set in the user's `.env`.\n4. **Use `document_id` from search results** `[id: uuid]` for get_document and list_versions.\n5. **Add metadata** -- at minimum `type` (\"decision-log\", \"research\", \"design-doc\") and `status` (\"active\", \"draft\").\n6. **Write structured Markdown** with H1/H2/H3 headings for good chunking and search.\n7. **Deletes are soft (recoverable); purge is web-UI-only.** If you decide to delete, surface it to the user (`I soft-deleted X — recoverable from the Cerefox web UI trash`). You cannot un-do your own delete from agent code by design.\n8. **Cross-doc links inside content**: **always use `[Text](document-uuid)`.** UUIDs are the only fully reliable link form — stable across title changes, never ambiguous, no encoding gotchas. Every `cerefox_search` result shows `[id: <uuid>]` after the title; grab it and use it. Title-based linking (`[Text](<Title With Spaces>)`) is fragile (breaks on colons, parens, ampersands, brackets — silently navigates to wrong page) — **don't write title-based links**; do an extra search to get the UUID instead. Repo-path forms (`[Text](docs/path.md)`) exist for repo-ingested files; don't construct manually. See `AGENT_GUIDE.md → Writing linkable content` for the full rule.\n9. **Concurrency: content updates require `expected_content_hash`.** Pass the `content_hash` you last saw — every read shows one (`cerefox_get_document` incl. outline mode, `cerefox_search`, `cerefox_metadata_search`) and **every write returns the new one, including create** (v1.3.0, #189), so after writing you already hold the token for your next edit; no re-read needed. If it's stale you get a **conflict** — re-read the document, merge your changes into the latest content, retry with the new hash. **Never resolve a conflict by overwriting blindly** — the current content includes another writer's work. `last_write_wins: true` skips the check; use it ONLY when an external source of truth makes conflicts meaningless (file re-sync), never to silence a conflict.\n10. **Search: prefer a few distinctive terms; heed `below confidence`.** When nothing clears the relevance threshold, `cerefox_search` returns the closest candidates prefixed with a `below confidence` warning instead of an empty set — that flag means **weak signal, not absent knowledge**: check the candidates' scores and titles before concluding the KB lacks the content. A truly empty response means nothing even weakly related exists.\n11. **Relations express how documents relate; lifecycle tells you if knowledge is still good.** Use `cerefox_set_relation` when one document supersedes, contradicts, references, or continues another. `supersedes` marks the target **superseded**; `contradicts` marks **both** stale; `related_to`/`duplicates`/`contradicts` are symmetric (both directions written). Any other type string is accepted without special behaviour. When a search result or `cerefox_get_relations` shows a neighbour marked `[superseded]` or `[stale]`, say so rather than presenting it as current.\n12. **Project memberships — non-destructive by default**: on `cerefox_ingest` updates, **`project_name` (singular) is a non-destructive add** (ensures membership, preserves others). Use **`project_names` (list)** when you want to set the doc's full project set in one call (destructive replace). For metadata-only project changes without writing content, use **`cerefox_set_document_projects(document_id, project_names)`** — that tool is the destructive-replace contract made explicit. Never call `cerefox_set_document_projects` with a single name when you mean \"add\" — that would REMOVE the doc from all other projects. When in doubt, use `cerefox_ingest` with singular `project_name`.\n\n## Update Workflow (ID-based -- preferred)\n\n```\nsearch(\"topic\") -> find doc [id: abc123] -> get_document(abc123) -> note its content_hash -> modify ->\ningest(title=\"Same Title\", content=\"...\", document_id=\"abc123\",\n       expected_content_hash=\"<the hash you read>\", author=\"my-agent\")\n```\n\nOn a **conflict** error: get_document again (fresh content + fresh hash) -> merge your changes -> retry with the new hash.\n\n## Update Workflow (title-based -- fallback)\n\n```\nsearch(\"topic\") -> find doc (note its hash) -> modify ->\ningest(title=\"Same Title\", content=\"...\", update_if_exists=true,\n       expected_content_hash=\"<the hash you read>\", author=\"my-agent\")\n```\n\n## Catch-Up Workflow\n\n```\nmetadata_search(metadata_filter={\"type\": \"decision-log\"}, updated_since=\"2026-03-28T00:00:00Z\")\n```\n\n## CLI fallback (when MCP is unavailable)\n\nIf `cerefox_search` is not in your tool list, your user has likely installed the Cerefox CLI. The canonical invocation is plain **`cerefox <subcommand>`** (the TypeScript CLI, installed via `npm install -g @cerefox/memory`). It uses a resource-verb shape (`cerefox document get`, `cerefox project list`, …).\n\nSame operations, same conventions. Full reference: [`docs/guides/cli.md`](docs/guides/cli.md). CLI flag names match MCP parameter names exactly (e.g. `metadata_filter` ↔ `--metadata-filter`); common flags also have single-letter short forms (`-f`, `-p`, `-c`, `-m`, `-u`, `-a`, `-r`). Use the canonical long name (what `--help` shows) or its short form — there are no long-form aliases like `--filter` or `--count`.\n\n| MCP tool | CLI |\n|---|---|\n| `cerefox_search` | `cerefox search \"<q>\" --requestor \"<your-name>\"` |\n| `cerefox_ingest` (paste) | `printf '...' \\| cerefox document ingest --paste --title \"<t>\" --author \"<your-name>\" --author-type agent` |\n| `cerefox_ingest` (update by ID) | `printf '...' \\| cerefox document ingest --paste --title \"<t>\" --document-id \"<uuid>\" --expected-content-hash \"<hash>\" --author \"<your-name>\" --author-type agent` |\n| `cerefox_get_document` | `cerefox document get <id> --version-id <vid> --requestor \"<your-name>\"` |\n| `cerefox_list_versions` | `cerefox document version list <id> --requestor \"<your-name>\"` |\n| `cerefox_list_projects` | `cerefox project list --requestor \"<your-name>\"` |\n| `cerefox_list_metadata_keys` | `cerefox metadata keys` |\n| `cerefox_set_relation` ⚑ | Link two documents (`source --rel_type--> target`) | `source_id`, `target_id`, `rel_type` (required), `metadata`, `author` |\n| `cerefox_delete_relation` ⚑ | Remove a relation | `source_id`, `target_id`, `rel_type` |\n| `cerefox_get_relations` ⚑ | All relations touching a document, both directions | `document_id` |\n| `cerefox_get_neighbors` ⚑ | Walk the graph along ONE relation type | `document_id`, `rel_type` (required), `depth`, `from_time`, `to_time`, `limit` |\n| `cerefox_metadata_search` | `cerefox metadata search --metadata-filter '<json>' --requestor \"<your-name>\"` (list a project: `cerefox document list --project <name>`) |\n| `cerefox_set_document_metadata` | `cerefox document set-metadata <id> --set key=value` (also `--remove key`, `--json '{...}'`, `--replace`) |\n| `cerefox_set_document_projects` | `cerefox document set-projects <id> <name...> --author \"<your-name>\" --author-type agent` (or `--clear` to remove all) |\n| `cerefox_get_audit_log` | `cerefox audit list --requestor \"<your-name>\"` (add `--json` for scripted access) |\n| `cerefox_get_help` | `cerefox guides show agent-quick-reference` (or `cerefox guides list` for the full bundled-docs index) |\n\n**Set identity on every call**, exactly as you would on MCP:\n- Writes (`document ingest`, `document ingest-dir`): `--author \"<your-name>\" --author-type agent`\n- Reads: `--requestor \"<your-name>\"`\n\nOr have your user set `CEREFOX_AUTHOR_NAME` / `CEREFOX_AUTHOR_TYPE` / `CEREFOX_REQUESTOR_NAME` in their `.env` to apply defaults once.\n\n## Timestamps are UTC\n\nEvery timestamp Cerefox returns — `created_at` on audit entries, version\nhistory, document metadata — is **UTC**, and now carries its `Z` marker so it\ncannot be mistaken for local time.\n\n**When you write a date into a document's CONTENT, use your own clock, not a\nCerefox timestamp.** These are different things: a timestamp records when the\nserver stored something; a date in a log entry or a heading is authored content\nand belongs to your timezone. An agent working a Pacific afternoon read\n`2026-08-11` from version history, wrote \"8/11\" into its entries, and put a\nday's work in the future — the timestamp was correct, and copying it into\ncontent was not.\n\nCerefox deliberately does not convert to local time on the API or MCP paths.\n\"Local\" has no server-side meaning: the remote MCP server runs in a cloud\nfunction whose local time *is* UTC, while a local MCP server runs in yours, so\nthe same document would report two different times depending on transport. The\nweb UI converts because a browser knows the viewer's timezone; nothing\nserver-side does.\n\n## Mistakes that have actually happened\n\nEach of these comes from a real agent session, and each is easy to make.\n\n- **`cerefox_ingest` always replaces the ENTIRE document.** Never a section.\n  Before sending, check that the tool name matches the intent: if the intent is\n  \"change one section\", the call is `cerefox_edit` with `replace_section`. A\n  section-sized edit sent as a full ingest truncated a 13,000-character index to\n  a single word. It was recovered from version history within the minute, but\n  only because it was noticed immediately.\n\n- **Do not include the anchor's own heading in your text.** `replace_section`\n  keeps the heading and `insert` places your text inside the section, so\n  including it produces two. This is now refused rather than silently applied,\n  but the shape is worth knowing: it happened twice in one session, the second\n  time while trying to repair the first. A *deeper* sub-heading inside your text\n  is fine.\n\n- **Content between sections belongs to the section ABOVE it.** A section runs\n  to the next heading of the same or higher level, so a `---` rule, a note, or\n  any trailing text sitting just above the next heading is part of the section\n  before it — even when it visually reads as belonging below. Replacing that\n  section takes it too. An agent hit exactly this: a `---` that separated two\n  major sections disappeared when the section above it was replaced. The write\n  was correct by the addressing rules; the surprise is that \"the end of this\n  section\" is further down the page than it looks. Note the loss warning will\n  not catch it if your replacement text is longer than what it replaced, since\n  there is then no net loss to report.\n\n- **To change only tags, use `cerefox_set_document_metadata`, never `cerefox_ingest`.**\n  Ingest replaces the whole document, so re-sending it to set one tag carries the\n  full transcription risk for no reason. The metadata tool merges: the keys you\n  pass are set, everything else is left alone, so you do not need to read the\n  document first and cannot drop a tag another agent set. Pass `null` as a value\n  to remove a key.\n\n- **Never partial-edit to fix a partial edit.** If a write leaves unexpected\n  structure, stop. Use `cerefox_list_versions`, retrieve the last good version,\n  and re-ingest cleanly. Repairing edits with more edits compounds the damage.\n\n- **A rejected batch is safe.** Operations in one `cerefox_edit` are\n  all-or-nothing: if any is invalid, nothing is written. A refusal costs you a\n  retry, not data — so prefer one call for changes that belong together, and do\n  not split a batch to \"make it more likely to succeed\".\n\n- **Read before replacing.** `cerefox_get_document(section: \"## Heading\")`\n  returns exactly what a `replace_section` on that anchor would overwrite. Use it\n  for any section you did not write in this session. The outline gives a\n  section's *size*, never its *text*.\n\n- **Verify after writing** — read the result back before reporting success, and\n  report what the read actually shows.\n\n- **Partial edits cannot change a document's stored TITLE.** `rename_section`\n  changes a heading inside the content; the title is a separate field and still\n  needs `cerefox_ingest`.\n\n- **If a capability seems missing from one server, suspect your client first.**\n  Local and remote run the same code. **Every `cerefox_get_help()` response\n  begins with the server's version and the operations it registers** — you do\n  not need a special topic, and the *absence* of that block is itself an answer:\n  a server that does not print it predates v1.5.0. If that\n  disagrees with your tool list, the client is holding a list it fetched before\n  an upgrade — clients cache it at connect time. Ask the user to restart the\n  client. Do not record a capability difference between servers as a fact; every\n  such report so far has been a stale client.\n";
+export const HELP_FULL = "# Cerefox Knowledge Base -- Agent Quick Reference\n\nCerefox is a persistent, shared knowledge base. You have **19 MCP tools** (18 of them have CLI equivalents — `cerefox_get_help` is MCP-only). For the full guide, search Cerefox for \"How AI Agents Use Cerefox\" or call `cerefox_get_help` to retrieve this content over MCP.\n\n## Tools\n\n| Tool | Purpose | Key params |\n|------|---------|------------|\n| `cerefox_search` | Find documents (hybrid FTS + semantic) | `query` (required), `project_name`, `metadata_filter`, `requestor` |\n| `cerefox_ingest` | Save or update a document | `title`, `content` (required), `document_id` (update by ID), `expected_content_hash` (**required on content updates** — see rule 9), `last_write_wins`, `update_if_exists`, `project_name` (single, non-destructive add on update), `project_names` (list, destructive replace on update), `metadata` (omit on update to keep existing tags; `{}` clears), `author` |\n| `cerefox_insert` | **Add** to a document without resending it. Cannot destroy content. | `document_id`, `text`, `position` (`end_of_document`/`end_of_section`/`after_heading`/`before_heading`), `expected_content_hash` (required), `anchor_heading` (unless `end_of_document`), `section_part` |\n| `cerefox_edit` | **Change** parts of a document: 1..n operations applied atomically | `document_id`, `operations` (`insert`/`replace_section`/`delete_section`/`rename_section`), `expected_content_hash` (required) |\n| `cerefox_delete_document` | **Soft**-delete a document (to trash; excluded from search; permanent purge is human-only) | `document_id`, `expected_content_hash` (**required** — a delete must follow a read), `reason` (recorded in the audit log — give one), `author` |\n| `cerefox_restore_document` | Restore a soft-deleted document from the trash (audited inverse of delete; no-op if not deleted) | `document_id` (required), `reason` (recorded in the audit log), `author` |\n| `cerefox_get_document` | Get full document by ID (header includes `content_hash` — the update token), or with `outline: true` just its heading paths, sizes and hash, or with `section: \"## Heading\"` one section's text | `document_id` (required), `outline`, `section`, `section_part` |\n| `cerefox_list_versions` | Version history of a document | `document_id` (required) |\n| `cerefox_set_relation` ⚑ | Link two documents (`source --rel_type--> target`) | `source_id`, `target_id`, `rel_type` (required), `metadata`, `author` |\n| `cerefox_delete_relation` ⚑ | Remove a relation | `source_id`, `target_id`, `rel_type` |\n| `cerefox_get_relations` ⚑ | All relations touching a document, both directions | `document_id` |\n| `cerefox_get_neighbors` ⚑ | Walk the graph along ONE relation type | `document_id`, `rel_type` (required), `depth`, `from_time`, `to_time`, `limit` |\n| `cerefox_metadata_search` | Find or list docs by metadata, project, or time (no text query) | `metadata_filter`, `project_name` (list a project's docs), `updated_since`, `include_content` — **at least one** of metadata_filter/project_name/updated_since/created_since |\n| `cerefox_list_metadata_keys` | Discover available metadata keys | (none required) |\n| `cerefox_list_projects` | List all projects | (none required) |\n| `cerefox_set_document_metadata` | Change tags WITHOUT resending content. **Merges** by default; a `null` value removes a key | `document_id`, `metadata` (required), `replace` (rare: set exactly this object), `author` |\n| `cerefox_set_document_projects` | Set doc's project memberships to exactly the given list (destructive replace; metadata-only, no content change) | `document_id`, `project_names` (required) |\n| `cerefox_get_audit_log` | Query write operation history | `document_id`, `author`, `operation`, `since` |\n| `cerefox_get_help` | Retrieve Cerefox conventions (this reference) over MCP. **Call this whenever uncertain.** | `topic` (optional, case-insensitive H2 substring match) |\n\n⚑ **Opt-in — usually absent.** The four relation tools are hidden unless the\noperator enables them (`relations_enabled`). **Trust your own tool list**: if\nthey are not in it, the feature is switched off for this deployment. That is\nnormal, not an error, and not something to work around.\n\n## Editing part of a document (prefer this over re-sending)\n\n**Re-sending a whole document to change part of it is the main way agents lose\ndata.** You have to reproduce the untouched remainder verbatim, and any drift\nsilently rewrites content nobody asked you to touch — which the caller cannot\ndiff. Use the partial-edit tools instead:\n\n1. **Learn the anchors** — `cerefox_get_document(document_id, outline: true)`.\n   Returns heading paths, per-section sizes and the `content_hash`, without the\n   body. The paths it returns are exactly what `anchor_heading` accepts.\n2. **Add** → `cerefox_insert`. `end_of_document` is a plain append;\n   `end_of_section` adds inside a named section. It is structurally incapable of\n   removing anything, so \"I meant to append\" cannot become \"I replaced the file\".\n3. **Look before you overwrite** — `cerefox_get_document(document_id,\n   section: \"## Heading\")` returns exactly the text a `replace_section` on that\n   anchor would destroy. The outline gives you a section's *size*, never its\n   *text*, so on a document you did not write yourself this is the difference\n   between a replace and a blind overwrite.\n4. **Change or remove** → `cerefox_edit`. Put changes that belong together in\n   ONE call: they apply atomically, so a table row and the total it feeds cannot\n   end up disagreeing. To change a single line, `replace_section` on its\n   smallest enclosing heading — that is the intended granularity, not a\n   workaround. To fix a stale heading (`## OPEN TODOs (as of ...)`), use\n   `rename_section`: it changes the heading text and leaves the body and\n   position alone.\n5. All of them require `expected_content_hash` and **have no last-write-wins**. A\n   conflict means someone else changed the document; re-read and decide, do not\n   force it.\n\n**A section runs to the next same-or-higher heading, or to the end of the\ndocument.** So `end_of_document` inserts land inside the *last* section, and\nreplacing or deleting that section removes them too. A large shrink in the\nresponse is your warning; `cerefox_list_versions` has the previous content.\n\n**When an anchor is ambiguous the tool refuses and hands you the options** — a\nrepeated heading returns the qualifying paths, and a section with both its own\ncontent and sub-sections returns both `section_part` choices. That is a\nrecoverable answer, not a failure: retry with what it gave you.\n\n## Essential Rules\n\n1. **Search before ingesting** -- check if the document exists first.\n2. **Prefer ID-based updates** -- pass `document_id` from search results for deterministic updates. Falls back to title-matching with `update_if_exists: true`.\n3. **Set `author`/`requestor`** to your name on every call (e.g., \"Claude Code\", \"archiver\"). On MCP, pass as parameters. On CLI, pass `--author`/`--author-type`/`--requestor` flags, or rely on `CEREFOX_AUTHOR_NAME`/`CEREFOX_AUTHOR_TYPE`/`CEREFOX_REQUESTOR_NAME` env vars set in the user's `.env`.\n4. **Use `document_id` from search results** `[id: uuid]` for get_document and list_versions.\n5. **Add metadata** -- at minimum `type` (\"decision-log\", \"research\", \"design-doc\") and `status` (\"active\", \"draft\").\n6. **Write structured Markdown** with H1/H2/H3 headings for good chunking and search.\n7. **Deletes are soft (recoverable); permanent purge is web-UI-only.** `cerefox_delete_document` requires the document's `content_hash` as you read it (read before you delete) and takes a `reason` — give one; it is what the human reviewing the trash sees. `cerefox_restore_document` undoes a mistaken delete (also audited, also takes a `reason`). Always surface deletes AND restores to the user. Once a human purges from the web UI, the document is gone for good.\n8. **Cross-doc links inside content**: **always use `[Text](document-uuid)`.** UUIDs are the only fully reliable link form — stable across title changes, never ambiguous, no encoding gotchas. Every `cerefox_search` result shows `[id: <uuid>]` after the title; grab it and use it. Title-based linking (`[Text](<Title With Spaces>)`) is fragile (breaks on colons, parens, ampersands, brackets — silently navigates to wrong page) — **don't write title-based links**; do an extra search to get the UUID instead. Repo-path forms (`[Text](docs/path.md)`) exist for repo-ingested files; don't construct manually. **The server validates `](uuid)` links on every write** (v1.7.0): a link to a nonexistent id rejects the write, naming the offender — that means you mangled the UUID; re-read the source and correct it, do not retry unchanged. Example ids go in backticks (code is not validated). `[[Wikilinks]]` may dangle. See `AGENT_GUIDE.md → Writing linkable content` for the full rule.\n9. **Concurrency: content updates require `expected_content_hash`.** Pass the `content_hash` you last saw — every read shows one (`cerefox_get_document` incl. outline mode, `cerefox_search`, `cerefox_metadata_search`) and **every write returns the new one, including create** (v1.3.0, #189), so after writing you already hold the token for your next edit; no re-read needed. If it's stale you get a **conflict** — re-read the document, merge your changes into the latest content, retry with the new hash. **Never resolve a conflict by overwriting blindly** — the current content includes another writer's work. `last_write_wins: true` skips the check; use it ONLY when an external source of truth makes conflicts meaningless (file re-sync), never to silence a conflict.\n10. **Search: prefer a few distinctive terms; heed `below confidence`.** When nothing clears the relevance threshold, `cerefox_search` returns the closest candidates prefixed with a `below confidence` warning instead of an empty set — that flag means **weak signal, not absent knowledge**: check the candidates' scores and titles before concluding the KB lacks the content. A truly empty response means nothing even weakly related exists.\n11. **Relations express how documents relate; lifecycle tells you if knowledge is still good.** Use `cerefox_set_relation` when one document supersedes, contradicts, references, or continues another. `supersedes` marks the target **superseded**; `contradicts` marks **both** stale; `related_to`/`duplicates`/`contradicts` are symmetric (both directions written). Any other type string is accepted without special behaviour. When a search result or `cerefox_get_relations` shows a neighbour marked `[superseded]` or `[stale]`, say so rather than presenting it as current.\n12. **Project memberships — non-destructive by default**: on `cerefox_ingest` updates, **`project_name` (singular) is a non-destructive add** (ensures membership, preserves others). Use **`project_names` (list)** when you want to set the doc's full project set in one call (destructive replace). For metadata-only project changes without writing content, use **`cerefox_set_document_projects(document_id, project_names)`** — that tool is the destructive-replace contract made explicit. Never call `cerefox_set_document_projects` with a single name when you mean \"add\" — that would REMOVE the doc from all other projects. When in doubt, use `cerefox_ingest` with singular `project_name`.\n\n## Update Workflow (ID-based -- preferred)\n\n```\nsearch(\"topic\") -> find doc [id: abc123] -> get_document(abc123) -> note its content_hash -> modify ->\ningest(title=\"Same Title\", content=\"...\", document_id=\"abc123\",\n       expected_content_hash=\"<the hash you read>\", author=\"my-agent\")\n```\n\nOn a **conflict** error: get_document again (fresh content + fresh hash) -> merge your changes -> retry with the new hash.\n\n## Update Workflow (title-based -- fallback)\n\n```\nsearch(\"topic\") -> find doc (note its hash) -> modify ->\ningest(title=\"Same Title\", content=\"...\", update_if_exists=true,\n       expected_content_hash=\"<the hash you read>\", author=\"my-agent\")\n```\n\n## Catch-Up Workflow\n\n```\nmetadata_search(metadata_filter={\"type\": \"decision-log\"}, updated_since=\"2026-03-28T00:00:00Z\")\n```\n\n## CLI fallback (when MCP is unavailable)\n\nIf `cerefox_search` is not in your tool list, your user has likely installed the Cerefox CLI. The canonical invocation is plain **`cerefox <subcommand>`** (the TypeScript CLI, installed via `npm install -g @cerefox/memory`). It uses a resource-verb shape (`cerefox document get`, `cerefox project list`, …).\n\nSame operations, same conventions. Full reference: [`docs/guides/cli.md`](docs/guides/cli.md). CLI flag names match MCP parameter names exactly (e.g. `metadata_filter` ↔ `--metadata-filter`); common flags also have single-letter short forms (`-f`, `-p`, `-c`, `-m`, `-u`, `-a`, `-r`). Use the canonical long name (what `--help` shows) or its short form — there are no long-form aliases like `--filter` or `--count`.\n\n| MCP tool | CLI |\n|---|---|\n| `cerefox_search` | `cerefox search \"<q>\" --requestor \"<your-name>\"` |\n| `cerefox_ingest` (paste) | `printf '...' \\| cerefox document ingest --paste --title \"<t>\" --author \"<your-name>\" --author-type agent` |\n| `cerefox_ingest` (update by ID) | `printf '...' \\| cerefox document ingest --paste --title \"<t>\" --document-id \"<uuid>\" --expected-content-hash \"<hash>\" --author \"<your-name>\" --author-type agent` |\n| `cerefox_get_document` | `cerefox document get <id> --version-id <vid> --requestor \"<your-name>\"` |\n| `cerefox_list_versions` | `cerefox document version list <id> --requestor \"<your-name>\"` |\n| `cerefox_list_projects` | `cerefox project list --requestor \"<your-name>\"` |\n| `cerefox_list_metadata_keys` | `cerefox metadata keys` |\n| `cerefox_insert` | `cerefox document insert <id> -t \"<text>\" -p <position> -a \"<anchor-heading>\" -e \"<hash>\" --requestor \"<your-name>\" --author-type agent` |\n| `cerefox_edit` | `cerefox document edit-parts <id> --operations '<json>' -e \"<hash>\" --requestor \"<your-name>\" --author-type agent` |\n| `cerefox_delete_document` | `cerefox document delete <id> --reason \"<why>\" --author \"<your-name>\" --author-type agent --yes` (confirms interactively instead of requiring the hash) |\n| `cerefox_restore_document` | `cerefox document restore <id> --reason \"<why>\" --author \"<your-name>\" --author-type agent` |\n| `cerefox_metadata_search` | `cerefox metadata search --metadata-filter '<json>' --requestor \"<your-name>\"` (list a project: `cerefox document list --project <name>`) |\n| `cerefox_set_document_metadata` | `cerefox document set-metadata <id> --set key=value` (also `--remove key`, `--json '{...}'`, `--replace`) |\n| `cerefox_set_document_projects` | `cerefox document set-projects <id> <name...> --author \"<your-name>\" --author-type agent` (or `--clear` to remove all) |\n| `cerefox_get_audit_log` | `cerefox audit list --requestor \"<your-name>\"` (add `--json` for scripted access) |\n| `cerefox_get_help` | `cerefox guides show agent-quick-reference` (or `cerefox guides list` for the full bundled-docs index) |\n\n**Set identity on every call**, exactly as you would on MCP:\n- Writes (`document ingest`, `document ingest-dir`): `--author \"<your-name>\" --author-type agent`\n- Reads: `--requestor \"<your-name>\"`\n\nOr have your user set `CEREFOX_AUTHOR_NAME` / `CEREFOX_AUTHOR_TYPE` / `CEREFOX_REQUESTOR_NAME` in their `.env` to apply defaults once.\n\n## Timestamps are UTC\n\nEvery timestamp Cerefox returns — `created_at` on audit entries, version\nhistory, document metadata — is **UTC**, and now carries its `Z` marker so it\ncannot be mistaken for local time.\n\n**When you write a date into a document's CONTENT, use your own clock, not a\nCerefox timestamp.** These are different things: a timestamp records when the\nserver stored something; a date in a log entry or a heading is authored content\nand belongs to your timezone. An agent working a Pacific afternoon read\n`2026-08-11` from version history, wrote \"8/11\" into its entries, and put a\nday's work in the future — the timestamp was correct, and copying it into\ncontent was not.\n\nCerefox deliberately does not convert to local time on the API or MCP paths.\n\"Local\" has no server-side meaning: the remote MCP server runs in a cloud\nfunction whose local time *is* UTC, while a local MCP server runs in yours, so\nthe same document would report two different times depending on transport. The\nweb UI converts because a browser knows the viewer's timezone; nothing\nserver-side does.\n\n## Mistakes that have actually happened\n\nEach of these comes from a real agent session, and each is easy to make.\n\n- **`cerefox_ingest` always replaces the ENTIRE document.** Never a section.\n  Before sending, check that the tool name matches the intent: if the intent is\n  \"change one section\", the call is `cerefox_edit` with `replace_section`. A\n  section-sized edit sent as a full ingest truncated a 13,000-character index to\n  a single word. It was recovered from version history within the minute, but\n  only because it was noticed immediately.\n\n- **Do not include the anchor's own heading in your text.** `replace_section`\n  keeps the heading and `insert` places your text inside the section, so\n  including it produces two. This is now refused rather than silently applied,\n  but the shape is worth knowing: it happened twice in one session, the second\n  time while trying to repair the first. A *deeper* sub-heading inside your text\n  is fine.\n\n- **Content between sections belongs to the section ABOVE it.** A section runs\n  to the next heading of the same or higher level, so a `---` rule, a note, or\n  any trailing text sitting just above the next heading is part of the section\n  before it — even when it visually reads as belonging below. Replacing that\n  section takes it too. An agent hit exactly this: a `---` that separated two\n  major sections disappeared when the section above it was replaced. The write\n  was correct by the addressing rules; the surprise is that \"the end of this\n  section\" is further down the page than it looks. Note the loss warning will\n  not catch it if your replacement text is longer than what it replaced, since\n  there is then no net loss to report.\n\n- **To change only tags, use `cerefox_set_document_metadata`, never `cerefox_ingest`.**\n  Ingest replaces the whole document, so re-sending it to set one tag carries the\n  full transcription risk for no reason. The metadata tool merges: the keys you\n  pass are set, everything else is left alone, so you do not need to read the\n  document first and cannot drop a tag another agent set. Pass `null` as a value\n  to remove a key.\n\n- **Never partial-edit to fix a partial edit.** If a write leaves unexpected\n  structure, stop. Use `cerefox_list_versions`, retrieve the last good version,\n  and re-ingest cleanly. Repairing edits with more edits compounds the damage.\n\n- **A rejected batch is safe.** Operations in one `cerefox_edit` are\n  all-or-nothing: if any is invalid, nothing is written. A refusal costs you a\n  retry, not data — so prefer one call for changes that belong together, and do\n  not split a batch to \"make it more likely to succeed\".\n\n- **Read before replacing.** `cerefox_get_document(section: \"## Heading\")`\n  returns exactly what a `replace_section` on that anchor would overwrite. Use it\n  for any section you did not write in this session. The outline gives a\n  section's *size*, never its *text*.\n\n- **Verify after writing** — read the result back before reporting success, and\n  report what the read actually shows.\n\n- **Partial edits cannot change a document's stored TITLE.** `rename_section`\n  changes a heading inside the content; the title is a separate field and still\n  needs `cerefox_ingest`.\n\n- **If a capability seems missing from one server, suspect your client first.**\n  Local and remote run the same code. **Every `cerefox_get_help()` response\n  begins with the server's version and the operations it registers** — you do\n  not need a special topic, and the *absence* of that block is itself an answer:\n  a server that does not print it predates v1.5.0. If that\n  disagrees with your tool list, the client is holding a list it fetched before\n  an upgrade — clients cache it at connect time. Ask the user to restart the\n  client. Do not record a capability difference between servers as a fact; every\n  such report so far has been a stale client.\n";
 /** Sections keyed by their H2 heading text (lower-cased for matching). */
 export const HELP_SECTIONS: Record<string, string> = {
-  "Tools": "## Tools\n\n| Tool | Purpose | Key params |\n|------|---------|------------|\n| `cerefox_search` | Find documents (hybrid FTS + semantic) | `query` (required), `project_name`, `metadata_filter`, `requestor` |\n| `cerefox_ingest` | Save or update a document | `title`, `content` (required), `document_id` (update by ID), `expected_content_hash` (**required on content updates** — see rule 9), `last_write_wins`, `update_if_exists`, `project_name` (single, non-destructive add on update), `project_names` (list, destructive replace on update), `metadata` (omit on update to keep existing tags; `{}` clears), `author` |\n| `cerefox_insert` | **Add** to a document without resending it. Cannot destroy content. | `document_id`, `text`, `position` (`end_of_document`/`end_of_section`/`after_heading`/`before_heading`), `expected_content_hash` (required), `anchor_heading` (unless `end_of_document`), `section_part` |\n| `cerefox_edit` | **Change** parts of a document: 1..n operations applied atomically | `document_id`, `operations` (`insert`/`replace_section`/`delete_section`/`rename_section`), `expected_content_hash` (required) |\n| `cerefox_get_document` | Get full document by ID (header includes `content_hash` — the update token), or with `outline: true` just its heading paths, sizes and hash, or with `section: \"## Heading\"` one section's text | `document_id` (required), `outline`, `section`, `section_part` |\n| `cerefox_list_versions` | Version history of a document | `document_id` (required) |\n| `cerefox_set_relation` ⚑ | Link two documents (`source --rel_type--> target`) | `source_id`, `target_id`, `rel_type` (required), `metadata`, `author` |\n| `cerefox_delete_relation` ⚑ | Remove a relation | `source_id`, `target_id`, `rel_type` |\n| `cerefox_get_relations` ⚑ | All relations touching a document, both directions | `document_id` |\n| `cerefox_get_neighbors` ⚑ | Walk the graph along ONE relation type | `document_id`, `rel_type` (required), `depth`, `from_time`, `to_time`, `limit` |\n| `cerefox_metadata_search` | Find or list docs by metadata, project, or time (no text query) | `metadata_filter`, `project_name` (list a project's docs), `updated_since`, `include_content` — **at least one** of metadata_filter/project_name/updated_since/created_since |\n| `cerefox_list_metadata_keys` | Discover available metadata keys | (none required) |\n| `cerefox_list_projects` | List all projects | (none required) |\n| `cerefox_set_document_metadata` | Change tags WITHOUT resending content. **Merges** by default; a `null` value removes a key | `document_id`, `metadata` (required), `replace` (rare: set exactly this object), `author` |\n| `cerefox_set_document_projects` | Set doc's project memberships to exactly the given list (destructive replace; metadata-only, no content change) | `document_id`, `project_names` (required) |\n| `cerefox_get_audit_log` | Query write operation history | `document_id`, `author`, `operation`, `since` |\n| `cerefox_get_help` | Retrieve Cerefox conventions (this reference) over MCP. **Call this whenever uncertain.** | `topic` (optional, case-insensitive H2 substring match) |\n\n⚑ **Opt-in — usually absent.** The four relation tools are hidden unless the\noperator enables them (`relations_enabled`). **Trust your own tool list**: if\nthey are not in it, the feature is switched off for this deployment. That is\nnormal, not an error, and not something to work around.",
+  "Tools": "## Tools\n\n| Tool | Purpose | Key params |\n|------|---------|------------|\n| `cerefox_search` | Find documents (hybrid FTS + semantic) | `query` (required), `project_name`, `metadata_filter`, `requestor` |\n| `cerefox_ingest` | Save or update a document | `title`, `content` (required), `document_id` (update by ID), `expected_content_hash` (**required on content updates** — see rule 9), `last_write_wins`, `update_if_exists`, `project_name` (single, non-destructive add on update), `project_names` (list, destructive replace on update), `metadata` (omit on update to keep existing tags; `{}` clears), `author` |\n| `cerefox_insert` | **Add** to a document without resending it. Cannot destroy content. | `document_id`, `text`, `position` (`end_of_document`/`end_of_section`/`after_heading`/`before_heading`), `expected_content_hash` (required), `anchor_heading` (unless `end_of_document`), `section_part` |\n| `cerefox_edit` | **Change** parts of a document: 1..n operations applied atomically | `document_id`, `operations` (`insert`/`replace_section`/`delete_section`/`rename_section`), `expected_content_hash` (required) |\n| `cerefox_delete_document` | **Soft**-delete a document (to trash; excluded from search; permanent purge is human-only) | `document_id`, `expected_content_hash` (**required** — a delete must follow a read), `reason` (recorded in the audit log — give one), `author` |\n| `cerefox_restore_document` | Restore a soft-deleted document from the trash (audited inverse of delete; no-op if not deleted) | `document_id` (required), `reason` (recorded in the audit log), `author` |\n| `cerefox_get_document` | Get full document by ID (header includes `content_hash` — the update token), or with `outline: true` just its heading paths, sizes and hash, or with `section: \"## Heading\"` one section's text | `document_id` (required), `outline`, `section`, `section_part` |\n| `cerefox_list_versions` | Version history of a document | `document_id` (required) |\n| `cerefox_set_relation` ⚑ | Link two documents (`source --rel_type--> target`) | `source_id`, `target_id`, `rel_type` (required), `metadata`, `author` |\n| `cerefox_delete_relation` ⚑ | Remove a relation | `source_id`, `target_id`, `rel_type` |\n| `cerefox_get_relations` ⚑ | All relations touching a document, both directions | `document_id` |\n| `cerefox_get_neighbors` ⚑ | Walk the graph along ONE relation type | `document_id`, `rel_type` (required), `depth`, `from_time`, `to_time`, `limit` |\n| `cerefox_metadata_search` | Find or list docs by metadata, project, or time (no text query) | `metadata_filter`, `project_name` (list a project's docs), `updated_since`, `include_content` — **at least one** of metadata_filter/project_name/updated_since/created_since |\n| `cerefox_list_metadata_keys` | Discover available metadata keys | (none required) |\n| `cerefox_list_projects` | List all projects | (none required) |\n| `cerefox_set_document_metadata` | Change tags WITHOUT resending content. **Merges** by default; a `null` value removes a key | `document_id`, `metadata` (required), `replace` (rare: set exactly this object), `author` |\n| `cerefox_set_document_projects` | Set doc's project memberships to exactly the given list (destructive replace; metadata-only, no content change) | `document_id`, `project_names` (required) |\n| `cerefox_get_audit_log` | Query write operation history | `document_id`, `author`, `operation`, `since` |\n| `cerefox_get_help` | Retrieve Cerefox conventions (this reference) over MCP. **Call this whenever uncertain.** | `topic` (optional, case-insensitive H2 substring match) |\n\n⚑ **Opt-in — usually absent.** The four relation tools are hidden unless the\noperator enables them (`relations_enabled`). **Trust your own tool list**: if\nthey are not in it, the feature is switched off for this deployment. That is\nnormal, not an error, and not something to work around.",
   "Editing part of a document (prefer this over re-sending)": "## Editing part of a document (prefer this over re-sending)\n\n**Re-sending a whole document to change part of it is the main way agents lose\ndata.** You have to reproduce the untouched remainder verbatim, and any drift\nsilently rewrites content nobody asked you to touch — which the caller cannot\ndiff. Use the partial-edit tools instead:\n\n1. **Learn the anchors** — `cerefox_get_document(document_id, outline: true)`.\n   Returns heading paths, per-section sizes and the `content_hash`, without the\n   body. The paths it returns are exactly what `anchor_heading` accepts.\n2. **Add** → `cerefox_insert`. `end_of_document` is a plain append;\n   `end_of_section` adds inside a named section. It is structurally incapable of\n   removing anything, so \"I meant to append\" cannot become \"I replaced the file\".\n3. **Look before you overwrite** — `cerefox_get_document(document_id,\n   section: \"## Heading\")` returns exactly the text a `replace_section` on that\n   anchor would destroy. The outline gives you a section's *size*, never its\n   *text*, so on a document you did not write yourself this is the difference\n   between a replace and a blind overwrite.\n4. **Change or remove** → `cerefox_edit`. Put changes that belong together in\n   ONE call: they apply atomically, so a table row and the total it feeds cannot\n   end up disagreeing. To change a single line, `replace_section` on its\n   smallest enclosing heading — that is the intended granularity, not a\n   workaround. To fix a stale heading (`## OPEN TODOs (as of ...)`), use\n   `rename_section`: it changes the heading text and leaves the body and\n   position alone.\n5. All of them require `expected_content_hash` and **have no last-write-wins**. A\n   conflict means someone else changed the document; re-read and decide, do not\n   force it.\n\n**A section runs to the next same-or-higher heading, or to the end of the\ndocument.** So `end_of_document` inserts land inside the *last* section, and\nreplacing or deleting that section removes them too. A large shrink in the\nresponse is your warning; `cerefox_list_versions` has the previous content.\n\n**When an anchor is ambiguous the tool refuses and hands you the options** — a\nrepeated heading returns the qualifying paths, and a section with both its own\ncontent and sub-sections returns both `section_part` choices. That is a\nrecoverable answer, not a failure: retry with what it gave you.",
-  "Essential Rules": "## Essential Rules\n\n1. **Search before ingesting** -- check if the document exists first.\n2. **Prefer ID-based updates** -- pass `document_id` from search results for deterministic updates. Falls back to title-matching with `update_if_exists: true`.\n3. **Set `author`/`requestor`** to your name on every call (e.g., \"Claude Code\", \"archiver\"). On MCP, pass as parameters. On CLI, pass `--author`/`--author-type`/`--requestor` flags, or rely on `CEREFOX_AUTHOR_NAME`/`CEREFOX_AUTHOR_TYPE`/`CEREFOX_REQUESTOR_NAME` env vars set in the user's `.env`.\n4. **Use `document_id` from search results** `[id: uuid]` for get_document and list_versions.\n5. **Add metadata** -- at minimum `type` (\"decision-log\", \"research\", \"design-doc\") and `status` (\"active\", \"draft\").\n6. **Write structured Markdown** with H1/H2/H3 headings for good chunking and search.\n7. **Deletes are soft (recoverable); purge is web-UI-only.** If you decide to delete, surface it to the user (`I soft-deleted X — recoverable from the Cerefox web UI trash`). You cannot un-do your own delete from agent code by design.\n8. **Cross-doc links inside content**: **always use `[Text](document-uuid)`.** UUIDs are the only fully reliable link form — stable across title changes, never ambiguous, no encoding gotchas. Every `cerefox_search` result shows `[id: <uuid>]` after the title; grab it and use it. Title-based linking (`[Text](<Title With Spaces>)`) is fragile (breaks on colons, parens, ampersands, brackets — silently navigates to wrong page) — **don't write title-based links**; do an extra search to get the UUID instead. Repo-path forms (`[Text](docs/path.md)`) exist for repo-ingested files; don't construct manually. See `AGENT_GUIDE.md → Writing linkable content` for the full rule.\n9. **Concurrency: content updates require `expected_content_hash`.** Pass the `content_hash` you last saw — every read shows one (`cerefox_get_document` incl. outline mode, `cerefox_search`, `cerefox_metadata_search`) and **every write returns the new one, including create** (v1.3.0, #189), so after writing you already hold the token for your next edit; no re-read needed. If it's stale you get a **conflict** — re-read the document, merge your changes into the latest content, retry with the new hash. **Never resolve a conflict by overwriting blindly** — the current content includes another writer's work. `last_write_wins: true` skips the check; use it ONLY when an external source of truth makes conflicts meaningless (file re-sync), never to silence a conflict.\n10. **Search: prefer a few distinctive terms; heed `below confidence`.** When nothing clears the relevance threshold, `cerefox_search` returns the closest candidates prefixed with a `below confidence` warning instead of an empty set — that flag means **weak signal, not absent knowledge**: check the candidates' scores and titles before concluding the KB lacks the content. A truly empty response means nothing even weakly related exists.\n11. **Relations express how documents relate; lifecycle tells you if knowledge is still good.** Use `cerefox_set_relation` when one document supersedes, contradicts, references, or continues another. `supersedes` marks the target **superseded**; `contradicts` marks **both** stale; `related_to`/`duplicates`/`contradicts` are symmetric (both directions written). Any other type string is accepted without special behaviour. When a search result or `cerefox_get_relations` shows a neighbour marked `[superseded]` or `[stale]`, say so rather than presenting it as current.\n12. **Project memberships — non-destructive by default**: on `cerefox_ingest` updates, **`project_name` (singular) is a non-destructive add** (ensures membership, preserves others). Use **`project_names` (list)** when you want to set the doc's full project set in one call (destructive replace). For metadata-only project changes without writing content, use **`cerefox_set_document_projects(document_id, project_names)`** — that tool is the destructive-replace contract made explicit. Never call `cerefox_set_document_projects` with a single name when you mean \"add\" — that would REMOVE the doc from all other projects. When in doubt, use `cerefox_ingest` with singular `project_name`.",
+  "Essential Rules": "## Essential Rules\n\n1. **Search before ingesting** -- check if the document exists first.\n2. **Prefer ID-based updates** -- pass `document_id` from search results for deterministic updates. Falls back to title-matching with `update_if_exists: true`.\n3. **Set `author`/`requestor`** to your name on every call (e.g., \"Claude Code\", \"archiver\"). On MCP, pass as parameters. On CLI, pass `--author`/`--author-type`/`--requestor` flags, or rely on `CEREFOX_AUTHOR_NAME`/`CEREFOX_AUTHOR_TYPE`/`CEREFOX_REQUESTOR_NAME` env vars set in the user's `.env`.\n4. **Use `document_id` from search results** `[id: uuid]` for get_document and list_versions.\n5. **Add metadata** -- at minimum `type` (\"decision-log\", \"research\", \"design-doc\") and `status` (\"active\", \"draft\").\n6. **Write structured Markdown** with H1/H2/H3 headings for good chunking and search.\n7. **Deletes are soft (recoverable); permanent purge is web-UI-only.** `cerefox_delete_document` requires the document's `content_hash` as you read it (read before you delete) and takes a `reason` — give one; it is what the human reviewing the trash sees. `cerefox_restore_document` undoes a mistaken delete (also audited, also takes a `reason`). Always surface deletes AND restores to the user. Once a human purges from the web UI, the document is gone for good.\n8. **Cross-doc links inside content**: **always use `[Text](document-uuid)`.** UUIDs are the only fully reliable link form — stable across title changes, never ambiguous, no encoding gotchas. Every `cerefox_search` result shows `[id: <uuid>]` after the title; grab it and use it. Title-based linking (`[Text](<Title With Spaces>)`) is fragile (breaks on colons, parens, ampersands, brackets — silently navigates to wrong page) — **don't write title-based links**; do an extra search to get the UUID instead. Repo-path forms (`[Text](docs/path.md)`) exist for repo-ingested files; don't construct manually. **The server validates `](uuid)` links on every write** (v1.7.0): a link to a nonexistent id rejects the write, naming the offender — that means you mangled the UUID; re-read the source and correct it, do not retry unchanged. Example ids go in backticks (code is not validated). `[[Wikilinks]]` may dangle. See `AGENT_GUIDE.md → Writing linkable content` for the full rule.\n9. **Concurrency: content updates require `expected_content_hash`.** Pass the `content_hash` you last saw — every read shows one (`cerefox_get_document` incl. outline mode, `cerefox_search`, `cerefox_metadata_search`) and **every write returns the new one, including create** (v1.3.0, #189), so after writing you already hold the token for your next edit; no re-read needed. If it's stale you get a **conflict** — re-read the document, merge your changes into the latest content, retry with the new hash. **Never resolve a conflict by overwriting blindly** — the current content includes another writer's work. `last_write_wins: true` skips the check; use it ONLY when an external source of truth makes conflicts meaningless (file re-sync), never to silence a conflict.\n10. **Search: prefer a few distinctive terms; heed `below confidence`.** When nothing clears the relevance threshold, `cerefox_search` returns the closest candidates prefixed with a `below confidence` warning instead of an empty set — that flag means **weak signal, not absent knowledge**: check the candidates' scores and titles before concluding the KB lacks the content. A truly empty response means nothing even weakly related exists.\n11. **Relations express how documents relate; lifecycle tells you if knowledge is still good.** Use `cerefox_set_relation` when one document supersedes, contradicts, references, or continues another. `supersedes` marks the target **superseded**; `contradicts` marks **both** stale; `related_to`/`duplicates`/`contradicts` are symmetric (both directions written). Any other type string is accepted without special behaviour. When a search result or `cerefox_get_relations` shows a neighbour marked `[superseded]` or `[stale]`, say so rather than presenting it as current.\n12. **Project memberships — non-destructive by default**: on `cerefox_ingest` updates, **`project_name` (singular) is a non-destructive add** (ensures membership, preserves others). Use **`project_names` (list)** when you want to set the doc's full project set in one call (destructive replace). For metadata-only project changes without writing content, use **`cerefox_set_document_projects(document_id, project_names)`** — that tool is the destructive-replace contract made explicit. Never call `cerefox_set_document_projects` with a single name when you mean \"add\" — that would REMOVE the doc from all other projects. When in doubt, use `cerefox_ingest` with singular `project_name`.",
   "Update Workflow (ID-based -- preferred)": "## Update Workflow (ID-based -- preferred)\n\n```\nsearch(\"topic\") -> find doc [id: abc123] -> get_document(abc123) -> note its content_hash -> modify ->\ningest(title=\"Same Title\", content=\"...\", document_id=\"abc123\",\n       expected_content_hash=\"<the hash you read>\", author=\"my-agent\")\n```\n\nOn a **conflict** error: get_document again (fresh content + fresh hash) -> merge your changes -> retry with the new hash.",
   "Update Workflow (title-based -- fallback)": "## Update Workflow (title-based -- fallback)\n\n```\nsearch(\"topic\") -> find doc (note its hash) -> modify ->\ningest(title=\"Same Title\", content=\"...\", update_if_exists=true,\n       expected_content_hash=\"<the hash you read>\", author=\"my-agent\")\n```",
   "Catch-Up Workflow": "## Catch-Up Workflow\n\n```\nmetadata_search(metadata_filter={\"type\": \"decision-log\"}, updated_since=\"2026-03-28T00:00:00Z\")\n```",
-  "CLI fallback (when MCP is unavailable)": "## CLI fallback (when MCP is unavailable)\n\nIf `cerefox_search` is not in your tool list, your user has likely installed the Cerefox CLI. The canonical invocation is plain **`cerefox <subcommand>`** (the TypeScript CLI, installed via `npm install -g @cerefox/memory`). It uses a resource-verb shape (`cerefox document get`, `cerefox project list`, …).\n\nSame operations, same conventions. Full reference: [`docs/guides/cli.md`](docs/guides/cli.md). CLI flag names match MCP parameter names exactly (e.g. `metadata_filter` ↔ `--metadata-filter`); common flags also have single-letter short forms (`-f`, `-p`, `-c`, `-m`, `-u`, `-a`, `-r`). Use the canonical long name (what `--help` shows) or its short form — there are no long-form aliases like `--filter` or `--count`.\n\n| MCP tool | CLI |\n|---|---|\n| `cerefox_search` | `cerefox search \"<q>\" --requestor \"<your-name>\"` |\n| `cerefox_ingest` (paste) | `printf '...' \\| cerefox document ingest --paste --title \"<t>\" --author \"<your-name>\" --author-type agent` |\n| `cerefox_ingest` (update by ID) | `printf '...' \\| cerefox document ingest --paste --title \"<t>\" --document-id \"<uuid>\" --expected-content-hash \"<hash>\" --author \"<your-name>\" --author-type agent` |\n| `cerefox_get_document` | `cerefox document get <id> --version-id <vid> --requestor \"<your-name>\"` |\n| `cerefox_list_versions` | `cerefox document version list <id> --requestor \"<your-name>\"` |\n| `cerefox_list_projects` | `cerefox project list --requestor \"<your-name>\"` |\n| `cerefox_list_metadata_keys` | `cerefox metadata keys` |\n| `cerefox_set_relation` ⚑ | Link two documents (`source --rel_type--> target`) | `source_id`, `target_id`, `rel_type` (required), `metadata`, `author` |\n| `cerefox_delete_relation` ⚑ | Remove a relation | `source_id`, `target_id`, `rel_type` |\n| `cerefox_get_relations` ⚑ | All relations touching a document, both directions | `document_id` |\n| `cerefox_get_neighbors` ⚑ | Walk the graph along ONE relation type | `document_id`, `rel_type` (required), `depth`, `from_time`, `to_time`, `limit` |\n| `cerefox_metadata_search` | `cerefox metadata search --metadata-filter '<json>' --requestor \"<your-name>\"` (list a project: `cerefox document list --project <name>`) |\n| `cerefox_set_document_metadata` | `cerefox document set-metadata <id> --set key=value` (also `--remove key`, `--json '{...}'`, `--replace`) |\n| `cerefox_set_document_projects` | `cerefox document set-projects <id> <name...> --author \"<your-name>\" --author-type agent` (or `--clear` to remove all) |\n| `cerefox_get_audit_log` | `cerefox audit list --requestor \"<your-name>\"` (add `--json` for scripted access) |\n| `cerefox_get_help` | `cerefox guides show agent-quick-reference` (or `cerefox guides list` for the full bundled-docs index) |\n\n**Set identity on every call**, exactly as you would on MCP:\n- Writes (`document ingest`, `document ingest-dir`): `--author \"<your-name>\" --author-type agent`\n- Reads: `--requestor \"<your-name>\"`\n\nOr have your user set `CEREFOX_AUTHOR_NAME` / `CEREFOX_AUTHOR_TYPE` / `CEREFOX_REQUESTOR_NAME` in their `.env` to apply defaults once.",
+  "CLI fallback (when MCP is unavailable)": "## CLI fallback (when MCP is unavailable)\n\nIf `cerefox_search` is not in your tool list, your user has likely installed the Cerefox CLI. The canonical invocation is plain **`cerefox <subcommand>`** (the TypeScript CLI, installed via `npm install -g @cerefox/memory`). It uses a resource-verb shape (`cerefox document get`, `cerefox project list`, …).\n\nSame operations, same conventions. Full reference: [`docs/guides/cli.md`](docs/guides/cli.md). CLI flag names match MCP parameter names exactly (e.g. `metadata_filter` ↔ `--metadata-filter`); common flags also have single-letter short forms (`-f`, `-p`, `-c`, `-m`, `-u`, `-a`, `-r`). Use the canonical long name (what `--help` shows) or its short form — there are no long-form aliases like `--filter` or `--count`.\n\n| MCP tool | CLI |\n|---|---|\n| `cerefox_search` | `cerefox search \"<q>\" --requestor \"<your-name>\"` |\n| `cerefox_ingest` (paste) | `printf '...' \\| cerefox document ingest --paste --title \"<t>\" --author \"<your-name>\" --author-type agent` |\n| `cerefox_ingest` (update by ID) | `printf '...' \\| cerefox document ingest --paste --title \"<t>\" --document-id \"<uuid>\" --expected-content-hash \"<hash>\" --author \"<your-name>\" --author-type agent` |\n| `cerefox_get_document` | `cerefox document get <id> --version-id <vid> --requestor \"<your-name>\"` |\n| `cerefox_list_versions` | `cerefox document version list <id> --requestor \"<your-name>\"` |\n| `cerefox_list_projects` | `cerefox project list --requestor \"<your-name>\"` |\n| `cerefox_list_metadata_keys` | `cerefox metadata keys` |\n| `cerefox_insert` | `cerefox document insert <id> -t \"<text>\" -p <position> -a \"<anchor-heading>\" -e \"<hash>\" --requestor \"<your-name>\" --author-type agent` |\n| `cerefox_edit` | `cerefox document edit-parts <id> --operations '<json>' -e \"<hash>\" --requestor \"<your-name>\" --author-type agent` |\n| `cerefox_delete_document` | `cerefox document delete <id> --reason \"<why>\" --author \"<your-name>\" --author-type agent --yes` (confirms interactively instead of requiring the hash) |\n| `cerefox_restore_document` | `cerefox document restore <id> --reason \"<why>\" --author \"<your-name>\" --author-type agent` |\n| `cerefox_metadata_search` | `cerefox metadata search --metadata-filter '<json>' --requestor \"<your-name>\"` (list a project: `cerefox document list --project <name>`) |\n| `cerefox_set_document_metadata` | `cerefox document set-metadata <id> --set key=value` (also `--remove key`, `--json '{...}'`, `--replace`) |\n| `cerefox_set_document_projects` | `cerefox document set-projects <id> <name...> --author \"<your-name>\" --author-type agent` (or `--clear` to remove all) |\n| `cerefox_get_audit_log` | `cerefox audit list --requestor \"<your-name>\"` (add `--json` for scripted access) |\n| `cerefox_get_help` | `cerefox guides show agent-quick-reference` (or `cerefox guides list` for the full bundled-docs index) |\n\n**Set identity on every call**, exactly as you would on MCP:\n- Writes (`document ingest`, `document ingest-dir`): `--author \"<your-name>\" --author-type agent`\n- Reads: `--requestor \"<your-name>\"`\n\nOr have your user set `CEREFOX_AUTHOR_NAME` / `CEREFOX_AUTHOR_TYPE` / `CEREFOX_REQUESTOR_NAME` in their `.env` to apply defaults once.",
   "Timestamps are UTC": "## Timestamps are UTC\n\nEvery timestamp Cerefox returns — `created_at` on audit entries, version\nhistory, document metadata — is **UTC**, and now carries its `Z` marker so it\ncannot be mistaken for local time.\n\n**When you write a date into a document's CONTENT, use your own clock, not a\nCerefox timestamp.** These are different things: a timestamp records when the\nserver stored something; a date in a log entry or a heading is authored content\nand belongs to your timezone. An agent working a Pacific afternoon read\n`2026-08-11` from version history, wrote \"8/11\" into its entries, and put a\nday's work in the future — the timestamp was correct, and copying it into\ncontent was not.\n\nCerefox deliberately does not convert to local time on the API or MCP paths.\n\"Local\" has no server-side meaning: the remote MCP server runs in a cloud\nfunction whose local time *is* UTC, while a local MCP server runs in yours, so\nthe same document would report two different times depending on transport. The\nweb UI converts because a browser knows the viewer's timezone; nothing\nserver-side does.",
   "Mistakes that have actually happened": "## Mistakes that have actually happened\n\nEach of these comes from a real agent session, and each is easy to make.\n\n- **`cerefox_ingest` always replaces the ENTIRE document.** Never a section.\n  Before sending, check that the tool name matches the intent: if the intent is\n  \"change one section\", the call is `cerefox_edit` with `replace_section`. A\n  section-sized edit sent as a full ingest truncated a 13,000-character index to\n  a single word. It was recovered from version history within the minute, but\n  only because it was noticed immediately.\n\n- **Do not include the anchor's own heading in your text.** `replace_section`\n  keeps the heading and `insert` places your text inside the section, so\n  including it produces two. This is now refused rather than silently applied,\n  but the shape is worth knowing: it happened twice in one session, the second\n  time while trying to repair the first. A *deeper* sub-heading inside your text\n  is fine.\n\n- **Content between sections belongs to the section ABOVE it.** A section runs\n  to the next heading of the same or higher level, so a `---` rule, a note, or\n  any trailing text sitting just above the next heading is part of the section\n  before it — even when it visually reads as belonging below. Replacing that\n  section takes it too. An agent hit exactly this: a `---` that separated two\n  major sections disappeared when the section above it was replaced. The write\n  was correct by the addressing rules; the surprise is that \"the end of this\n  section\" is further down the page than it looks. Note the loss warning will\n  not catch it if your replacement text is longer than what it replaced, since\n  there is then no net loss to report.\n\n- **To change only tags, use `cerefox_set_document_metadata`, never `cerefox_ingest`.**\n  Ingest replaces the whole document, so re-sending it to set one tag carries the\n  full transcription risk for no reason. The metadata tool merges: the keys you\n  pass are set, everything else is left alone, so you do not need to read the\n  document first and cannot drop a tag another agent set. Pass `null` as a value\n  to remove a key.\n\n- **Never partial-edit to fix a partial edit.** If a write leaves unexpected\n  structure, stop. Use `cerefox_list_versions`, retrieve the last good version,\n  and re-ingest cleanly. Repairing edits with more edits compounds the damage.\n\n- **A rejected batch is safe.** Operations in one `cerefox_edit` are\n  all-or-nothing: if any is invalid, nothing is written. A refusal costs you a\n  retry, not data — so prefer one call for changes that belong together, and do\n  not split a batch to \"make it more likely to succeed\".\n\n- **Read before replacing.** `cerefox_get_document(section: \"## Heading\")`\n  returns exactly what a `replace_section` on that anchor would overwrite. Use it\n  for any section you did not write in this session. The outline gives a\n  section's *size*, never its *text*.\n\n- **Verify after writing** — read the result back before reporting success, and\n  report what the read actually shows.\n\n- **Partial edits cannot change a document's stored TITLE.** `rename_section`\n  changes a heading inside the content; the title is a separate field and still\n  needs `cerefox_ingest`.\n\n- **If a capability seems missing from one server, suspect your client first.**\n  Local and remote run the same code. **Every `cerefox_get_help()` response\n  begins with the server's version and the operations it registers** — you do\n  not need a special topic, and the *absence* of that block is itself an answer:\n  a server that does not print it predates v1.5.0. If that\n  disagrees with your tool list, the client is holding a list it fetched before\n  an upgrade — clients cache it at connect time. Ask the user to restart the\n  client. Do not record a capability difference between servers as a fact; every\n  such report so far has been a stale client.",
 };

package/dist/server-assets/_shared/mcp-tools/index.ts CHANGED Viewed

@@ -8,10 +8,12 @@
  * Adding a new tool: write `<name>.ts` exporting a `ToolDefinition`, then
  * add it to `ALL_TOOLS` below. Both consumers pick it up automatically.
  *
- * Tool surface: 16 tools (12 core + 4 dormant relation tools).
+ * Tool surface: 19 tools (15 core + 4 dormant relation tools).
  */
 import { auditLogTool } from "./audit-log.ts";
+import { deleteDocumentTool } from "./delete-document.ts";
+import { restoreDocumentTool } from "./restore-document.ts";
 import {
   disabledToolMessage,
   relationsEnabled,
@@ -45,6 +47,11 @@ export const ALL_TOOLS: ToolDefinition[] = [
   // a client can grant it freely; edit carries the destructive operations.
   insertTool,
   editTool,
+  // Soft delete + restore (#208, #210): the content-lifecycle pair. Delete
+  // requires the caller's read-hash; restore is its audited inverse. Only
+  // permanent purge stays web-UI-only (trust model).
+  deleteDocumentTool,
+  restoreDocumentTool,
   getDocumentTool,
   listVersionsTool,
   metadataSearchTool,

package/dist/server-assets/_shared/mcp-tools/ingest.ts CHANGED Viewed

@@ -26,7 +26,7 @@ import {
 } from "./_chunker.ts";
 import { activeEmbedderName, embedBatch, resolveEmbedderKind } from "../embeddings/index.ts";
 import { ensureDocumentInProject, setDocumentProjectsByName } from "./_projects.ts";
-import { logUsage } from "./_utils.ts";
+import { extractConflictHashes, logUsage } from "./_utils.ts";
 import { McpInvalidParams, type ToolContext, type ToolDefinition } from "./types.ts";
 /**
@@ -48,10 +48,27 @@ function conflictError(documentId: string, expectedHash: string, currentHash: st
 /** Map RPC-side CEREFOX_CONFLICT / CEREFOX_TOKEN_REQUIRED errors to agent-first text. */
 function mapIngestRpcError(message: string, documentId: string): Error {
   if (message.includes("CEREFOX_CONFLICT")) {
-    const current = message.match(/current hash ([0-9a-f]{64})/)?.[1] ?? "unknown";
-    const expected = message.match(/expected hash ([0-9a-f]{64})/)?.[1] ?? "unknown";
+    const { expected, current } = extractConflictHashes(message);
     return conflictError(documentId, expected, current);
   }
+  if (message.includes("CEREFOX_UNRESOLVED_LINKS")) {
+    const ids = message.match(/do not exist: ([^.]+)\./)?.[1] ?? "(unparsed)";
+    return new Error(
+      `Write rejected — the content links document id(s) that do not exist: ${ids}. ` +
+        `These are almost certainly mangled UUIDs (long random ids corrupt easily when ` +
+        `regenerated). Do NOT retry unchanged: re-read the SOURCE you copied each link ` +
+        `from, correct the id(s), and resend. If an id is a deliberate example rather ` +
+        `than a real link, put it in code formatting (backticks or a fence).`,
+    );
+  }
+  if (message.includes("CEREFOX_DELETED")) {
+    const id = message.match(/document ([0-9a-f-]{36})/)?.[1] ?? documentId;
+    return new Error(
+      `Document ${id} is soft-deleted (in the trash). A trashed document cannot be ` +
+        `updated — restore it first with cerefox_restore_document, then retry, or ` +
+        `create a new document instead.`,
+    );
+  }
   if (message.includes("CEREFOX_TOKEN_REQUIRED")) {
     const current = message.match(/Current hash: ([0-9a-f]{64})/)?.[1];
     return new Error(
@@ -122,11 +139,13 @@ async function handler(
   // ── ID-based update path ─────────────────────────────────────────────────
   if (document_id) {
+    // No deleted_at filter: a trashed document must be DISTINGUISHABLE from a
+    // missing one — "not found" for a doc sitting in the trash sends the
+    // agent hunting for a typo instead of at the actual remedy.
     const { data: existing } = await supabase
       .from("cerefox_documents")
-      .select("id, title, content_hash")
+      .select("id, title, content_hash, deleted_at")
       .eq("id", document_id)
-      .is("deleted_at", null)
       .limit(1);
     if (!existing?.length) {
@@ -135,6 +154,16 @@ async function handler(
     const existingDoc = existing[0];
+    if (existingDoc.deleted_at) {
+      // Fast-fail before the embedding spend; the authoritative guard is in
+      // the RPC (0.12.0).
+      throw new McpInvalidParams(
+        `Document ${document_id} ("${existingDoc.title}") is soft-deleted (in the trash). ` +
+          `A trashed document cannot be updated — restore it first with ` +
+          `cerefox_restore_document, then retry, or create a new document instead.`,
+      );
+    }
     if (existingDoc.content_hash === contentHash) {
       const note = update_if_exists
         ? ""
@@ -206,16 +235,42 @@ async function handler(
   // ── Update-existing path ─────────────────────────────────────────────────
   if (update_if_exists) {
-    const { data: existing } = await supabase
+    // Prefer-live: a live title match always wins over a trashed twin.
+    // Recency alone would pick a freshly-trashed doc (soft delete bumps
+    // updated_at) and make the live document unreachable via update_if_exists.
+    let { data: existing } = await supabase
       .from("cerefox_documents")
-      .select("id, title, content_hash")
+      .select("id, title, content_hash, deleted_at")
       .eq("title", title)
+      .is("deleted_at", null)
       .order("updated_at", { ascending: false })
       .limit(1);
+    if (!existing?.length) {
+      ({ data: existing } = await supabase
+        .from("cerefox_documents")
+        .select("id, title, content_hash, deleted_at")
+        .eq("title", title)
+        .order("updated_at", { ascending: false })
+        .limit(1));
+    }
     if (existing?.length) {
       const existingDoc = existing[0];
+      if (existingDoc.deleted_at) {
+        // The title matched a TRASHED document (re-ingest after a delete —
+        // the "I forgot I deleted fotis.md" workflow). Refuse with the
+        // remedy, before the embedding spend: updating it would write into
+        // a search-excluded document, and creating a twin would confuse
+        // every later title match.
+        throw new McpInvalidParams(
+          `A document titled "${existingDoc.title}" is in the trash (soft-deleted, ` +
+            `id: ${existingDoc.id}). Restore it first with cerefox_restore_document ` +
+            `and re-ingest to update it — or have your user purge it from the web UI ` +
+            `Trash to start fresh.`,
+        );
+      }
       if (existingDoc.content_hash === contentHash) {
         return `Document already up-to-date: "${existingDoc.title}" (id: ${existingDoc.id}). Content hash unchanged (${contentHash}).`;
       }
@@ -282,13 +337,25 @@ async function handler(
   }
   // ── Hash deduplication (create path) ─────────────────────────────────────
+  // content_hash is unique STORE-WIDE, trashed documents included, so this
+  // check deliberately sees them: an insert would collide either way.
   const { data: hashMatch } = await supabase
     .from("cerefox_documents")
-    .select("id, title")
+    .select("id, title, deleted_at")
     .eq("content_hash", contentHash)
     .limit(1);
   if (hashMatch?.length) {
+    if (hashMatch[0].deleted_at) {
+      // "Already up-to-date" while search shows nothing is the classic
+      // re-upload-after-delete confusion. Say where the content actually is.
+      return (
+        `This exact content is already in the TRASH as "${hashMatch[0].title}" ` +
+        `(soft-deleted, id: ${hashMatch[0].id}). Restore it with ` +
+        `cerefox_restore_document instead of re-ingesting — or have your user ` +
+        `purge it from the web UI Trash to start fresh.`
+      );
+    }
     return `Document already up-to-date: "${hashMatch[0].title}" (id: ${hashMatch[0].id}). Content hash unchanged (${contentHash}) — pass it as expected_content_hash to edit it.`;
   }