docs: remove get_schema and research_web tools and the schema-and-web page

Delete the Schema & web tool reference page and strip references to
get_schema and research_web from the tool overview, three how-to
guides, and the bundled skill prompt. Retitle the skill's section I to
"Raw SQL (advanced fallback)" since only execute_query remains.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

Author: Aditya Thiruvengadam

docs/docs.json

@@ -93,8 +93,7 @@
               "tools/artifacts",
               "tools/context-graph",
               "tools/objectives",
-              "tools/workspace",
-              "tools/schema-and-web"
+              "tools/workspace"
             ]
           }
         ]

docs/guides/code-and-agents.mdx

@@ -34,9 +34,6 @@ In Claude Code, Cursor, or Codex, the assistant can chain process intelligence w
   <Step title="Confirm observation shape">
     Call [`list_process_observations`](/tools/observations-and-changes) and [`get_observation_activity_timeline`](/tools/observations-and-changes) on a sample observation. The assistant now knows the real fields and timing semantics.
   </Step>
-  <Step title="Pull schema if the code touches the DB">
-    Call [`get_schema`](/tools/schema-and-web) for the full PostgreSQL schema. Treat it as last-resort context — prefer typed tools where they exist.
-  </Step>
   <Step title="Generate, test, iterate">
     Standard code-gen loop, but every assumption is checkable against a live MCP call.
   </Step>

docs/guides/deep-research.mdx

@@ -28,7 +28,7 @@ The pattern: **Klarity tells the assistant what you actually do. Web research te
     Call [`get_process_observations`](/tools/observations-and-changes) and [`get_observation_activity_timeline`](/tools/observations-and-changes) — your real deviations, not assumed ones.
   </Step>
   <Step title="Pull external context">
-    Use the client's web search (or [`research_web`](/tools/schema-and-web) if you need it inside the MCP loop) to gather published frameworks, vendor positioning, or regulatory guidance.
+    Use the client's web search to gather published frameworks, vendor positioning, or regulatory guidance.
   </Step>
   <Step title="Synthesize the gap">
     "Your process does X today, with these deviations. The framework recommends Y. The gap is Z — here's the evidence for both sides."

docs/guides/find-opportunities.mdx

@@ -47,9 +47,6 @@ A platform lead, AI architect, or transformation owner asks where to focus next.
   <Step title="Synthesize a ranked candidate set">
     Process IDs, observation counts, dependency depth, similarity to known patterns, evidence trail.
   </Step>
-  <Step title="Cross-check vendor tooling (optional)">
-    Call [`research_web`](/tools/schema-and-web) on the highest-ranked candidates to see if there's published vendor or framework guidance worth pulling in.
-  </Step>
 </Steps>
 
 ## What to surface in the brief

docs/tools/overview.mdx

@@ -20,7 +20,6 @@ Every workflow reduces to: **find the right process(es) → fetch detail → gat
 | [Context Graph](/tools/context-graph) | Entities, communities, relationships, lineage | The question is relational |
 | [Objectives](/tools/objectives) | Advisor objective findings, actions, agent state | The user is continuing in-flight transformation work |
 | [Workspace](/tools/workspace) | List workspaces, switch active workspace, attribute configs | Orienting before drilling in, or switching context |
-| [Schema & web](/tools/schema-and-web) | Database schema, raw SQL, web research | Last-resort fallback for analytics or external context |
 
 ## Before you call
 

docs/tools/schema-and-web.mdx

@@ -163,24 +163,15 @@ build on prior analysis instead of starting from scratch.
 
 ---
 
-## I. Database & schema (advanced fallback)
+## I. Raw SQL (advanced fallback)
 
 | Tool | When to use |
 |---|---|
-| `get_schema` | Read the workspace database schema. |
 | `execute_query` | Raw SQL against the workspace DB. **Last resort** — prefer the typed tools above. Useful when the customer asks an analytics question no typed tool covers. |
 
 ---
 
-## J. Web research
-
-| Tool | When to use |
-|---|---|
-| `research_web` | Research a topic on the public web. Use sparingly — Klarity's value is grounded in the customer's workspace, not the open web. |
-
----
-
-## K. Workspace attributes & sessions
+## J. Workspace attributes & sessions
 
 | Tool | When to use |
 |---|---|
@@ -326,8 +317,6 @@ hierarchy.
    - `get_downstream_dependencies` — first cut at blast radius.
 10. Synthesize a ranked candidate set: process IDs, observation counts,
     dependency depth, similarity to known patterns, evidence trail.
-11. Optional: `research_web` to cross-check vendor automation tooling
-    for the highest-ranked candidates.
 
 ### Scenario 4 — Form a transformation thesis on a chosen process or value stream