t27mcp: examples.t27 copies find and has from check_hints.t27 (duplicate-bodies) #10166

Problem

specs/t27mcp/check_hints.t27 and specs/t27mcp/examples.t27 carry two copies of the same string helpers, find and has. The duplicate-bodies gate (tools/dupe_scan.py) rejects this on #10157: "a body written under find is now copied [2] time(s) and was in no ledger", and the same for has.

Fix

examples.t27 takes find and has from check_hints.t27 with use t27mcp::check_hints; and keeps no copy. The t27 MCP server (gHashTag/t27-mcp) gets its wasm/examples.wasm and wasm/examples_table.mjs regenerated from the new spec.

Done when

• python3 tools/dupe_scan.py reports no new duplicate body.
• t27c test-report specs/t27mcp/examples.t27: 8 of 8 pass, 0 vacuous; the negative control (a row widened past 40 lines) still fails.
• The regenerated wasm answers the same as the old one for topic_code and pick.
• gHashTag/t27-mcp npm run smoke passes.

Part of the epic #10066.

Boundary

• specs/t27mcp/examples.t27
• gHashTag/t27-mcp wasm/examples.wasm
• gHashTag/t27-mcp wasm/examples_table.mjs
• gHashTag/t27-mcp wasm/BUILD

User Scenarios

• Given a PR that adds specs/t27mcp/examples.t27, when the duplicate-bodies gate runs, then it finds no new copy of find or has.
• Given the question "clocked counter with sync reset", when the MCP server calls topic_code, then it gets the same topic as before the change.

Requirements

• FR-001: examples.t27 MUST NOT define find, has or NO_POS; it MUST import them with use t27mcp::check_hints;.
• FR-002: The regenerated examples.wasm MUST return the same topic_code and pick values as the one it replaces.
• FR-003: wasm/BUILD MUST carry the new sha256 of the spec, the table and the wasm.