diff --git a/workers/src/index.ts b/workers/src/index.ts
index 1ac23a6..999e3da 100644
--- a/workers/src/index.ts
+++ b/workers/src/index.ts
@@ -22,12 +22,57 @@ import { ZipBaselineFetcher } from "./zip-baseline-fetcher";
 import { RequestTracer } from "./tracing";
 import { parseConsumerLabel } from "./telemetry";
 import { renderNotFoundPage } from "./not-found-ui";
+import { parseTableRow } from "./markdown-utils";
 import pkg from "../package.json";
 
 export type { Env };
 
 const BUILD_VERSION = pkg.version;
 
+// ──────────────────────────────────────────────────────────────────────────────
+// Canon-table parsing helper.
+//
+// parseSelfReportHeadersTable extracts the self-report header contract from
+// canon/constraints/telemetry-governance.md. The table format is governed by
+// the canon doc itself; this parser is deliberately permissive (whitespace,
+// backticks around header name) and fails closed to null so the caller can
+// fall back to the minimal baseline without hiding the degradation.
+// ──────────────────────────────────────────────────────────────────────────────
+
+function parseSelfReportHeadersTable(markdown: string): Record<string, string> | null {
+  // Target section: "### Self-Report Fields" — grab the table that follows.
+  // Stop at the next `###` or `##` heading, whichever comes first.
+  //
+  // Expected table schema (governed by canon/constraints/telemetry-governance):
+  //   | Field | Header | Source | Description |
+  //   cols[0]  cols[1]  cols[2]  cols[3]
+  //
+  // We key on the Header (col 1, with backticks stripped) and use the
+  // Description (col 3) as the value. The parser is deliberately permissive
+  // on whitespace and fails closed to null so the caller falls back to the
+  // minimal baseline rather than hiding the degradation.
+  const section = markdown.match(
+    /###\s+Self-Report Fields[^\n]*\n([\s\S]*?)(?=\n###|\n##|$)/,
+  );
+  if (!section) return null;
+
+  const headers: Record<string, string> = {};
+  for (const raw of section[1].split("\n")) {
+    if (!raw.includes("|")) continue;
+    const cols = parseTableRow(raw);
+    // Need at least 4 cols (Field, Header, Source, Description).
+    // Skip header row, separator row, and any malformed row.
+    if (cols.length < 4) continue;
+    const headerName = cols[1].replace(/`/g, "").trim();
+    if (!headerName.startsWith("x-oddkit-")) continue; // skip header/separator
+    const description = cols[3].trim();
+    if (!description) continue;
+    headers[headerName] = description;
+  }
+
+  return Object.keys(headers).length > 0 ? headers : null;
+}
+
 // ──────────────────────────────────────────────────────────────────────────────
 // Consumer identification nudge
 //
@@ -451,7 +496,7 @@ Time filter example: WHERE timestamp > NOW() - INTERVAL '30' DAY`,
 
   server.tool(
     "telemetry_policy",
-    "Return oddkit telemetry and sharing policy guidance. What is tracked, what is excluded, and why. Fetched from canonical governance document at runtime.",
+    "Return oddkit telemetry and sharing policy guidance. What is tracked, what is excluded, and why. Fetched from canonical governance document at runtime. Response envelope declares governance_source (canon|baseline|minimal) per canon/constraints/core-governance-baseline.",
     {},
     {
       readOnlyHint: true,
@@ -460,15 +505,50 @@ Time filter example: WHERE timestamp > NOW() - INTERVAL '30' DAY`,
       openWorldHint: true,
     },
     async () => {
-      // Fetch the governance doc from canon
+      // Governance resolution per canon/constraints/core-governance-baseline:
+      //   1. Live canon fetch (preferred) → governance_source: "canon"
+      //   2. Minimal baseline (shipped in code) → governance_source: "minimal"
+      //
+      // This canary refactor implements tiers 1 and 3 only. The bundled
+      // baseline tier (2) and the build-time schema check arrive in follow-up
+      // work; the manifest + baseline directory are not yet in place.
       const fetcher = new ZipBaselineFetcher(env);
-      let policyContent = "Governance document not found. See https://github.com/klappy/klappy.dev/blob/main/canon/constraints/telemetry-governance.md";
+      let policyContent: string | null = null;
+      let selfReportHeaders: Record<string, string> | null = null;
+      let governanceSource: "canon" | "baseline" | "minimal" = "minimal";
 
       try {
         const content = await fetcher.getFile("canon/constraints/telemetry-governance.md");
-        if (content) policyContent = content;
+        if (content) {
+          policyContent = content;
+          const parsed = parseSelfReportHeadersTable(content);
+          if (parsed && Object.keys(parsed).length > 0) {
+            selfReportHeaders = parsed;
+            governanceSource = "canon";
+          }
+        }
       } catch {
-        // Fall through to default message
+        // Fall through to minimal tier below
+      }
+
+      if (governanceSource === "minimal") {
+        // Minimal baseline — the tool remains useful when canon is unreachable
+        // or the table cannot be parsed. These eight headers are the stable
+        // self-report contract; if canon adds a 9th, the "canon" tier delivers
+        // it and this list stays as the floor.
+        selfReportHeaders = {
+          "x-oddkit-client": "Your client name (highest priority identifier)",
+          "x-oddkit-client-version": "Your client version",
+          "x-oddkit-agent-name": "The AI agent name",
+          "x-oddkit-agent-version": "The AI agent version",
+          "x-oddkit-surface": "Where this is running (e.g. claude.ai, vscode)",
+          "x-oddkit-contact-url": "URL for your project or org",
+          "x-oddkit-policy-url": "Your privacy/telemetry policy URL",
+          "x-oddkit-capabilities": "Comma-separated capability list",
+        };
+        if (!policyContent) {
+          policyContent = "Governance document not reachable. See https://github.com/klappy/klappy.dev/blob/main/canon/constraints/telemetry-governance.md";
+        }
       }
 
       return {
@@ -479,16 +559,8 @@ Time filter example: WHERE timestamp > NOW() - INTERVAL '30' DAY`,
             result: {
               policy: policyContent,
               governance_uri: "klappy://canon/constraints/telemetry-governance",
-              self_report_headers: {
-                "x-oddkit-client": "Your client name (highest priority identifier)",
-                "x-oddkit-client-version": "Your client version",
-                "x-oddkit-agent-name": "The AI agent name",
-                "x-oddkit-agent-version": "The AI agent version",
-                "x-oddkit-surface": "Where this is running (e.g. claude.ai, vscode)",
-                "x-oddkit-contact-url": "URL for your project or org",
-                "x-oddkit-policy-url": "Your privacy/telemetry policy URL",
-                "x-oddkit-capabilities": "Comma-separated capability list",
-              },
+              governance_source: governanceSource,
+              self_report_headers: selfReportHeaders,
               generated_at: new Date().toISOString(),
             },
           }, null, 2),
diff --git a/workers/src/markdown-utils.ts b/workers/src/markdown-utils.ts
new file mode 100644
index 0000000..4158202
--- /dev/null
+++ b/workers/src/markdown-utils.ts
@@ -0,0 +1,24 @@
+/**
+ * Shared markdown parsing helpers.
+ *
+ * Keep this module dependency-free so it can be imported from any code path
+ * (orchestrate, index, future canon readers) without pulling in unrelated
+ * state. Every helper here must be pure and stateless.
+ */
+
+/**
+ * Parse a single markdown table row into trimmed cell values, preserving
+ * legitimately-empty middle cells. Only the leading and trailing empty strings
+ * produced by splitting a `| a | b |`-style row are stripped — a prior
+ * `.filter(c => c.length > 0)` approach also dropped empty interior cells,
+ * which silently collapsed the column count and caused `cols.length >= N`
+ * guards to misfire (e.g. a voice-dump row with an empty tiers cell).
+ */
+export function parseTableRow(row: string): string[] {
+  const parts = row.split("|");
+  // Strip the leading empty produced by a leading `|`, if present
+  if (parts.length > 0 && parts[0].trim() === "") parts.shift();
+  // Strip the trailing empty produced by a trailing `|`, if present
+  if (parts.length > 0 && parts[parts.length - 1].trim() === "") parts.pop();
+  return parts.map((c) => c.trim());
+}
diff --git a/workers/src/orchestrate.ts b/workers/src/orchestrate.ts
index 401529a..d50ba86 100644
--- a/workers/src/orchestrate.ts
+++ b/workers/src/orchestrate.ts
@@ -18,6 +18,7 @@ import {
   type SectionResult,
 } from "./zip-baseline-fetcher";
 import { buildBM25Index, searchBM25, type BM25Index } from "./bm25";
+import { parseTableRow } from "./markdown-utils";
 import type { RequestTracer } from "./tracing";
 import pkg from "../package.json";
 
@@ -154,27 +155,6 @@ export interface OrchestrateOptions {
   canonUrl?: string;
 }
 
-// ──────────────────────────────────────────────────────────────────────────────
-// Markdown table helpers
-// ──────────────────────────────────────────────────────────────────────────────
-
-/**
- * Parse a single markdown table row into trimmed cell values, preserving
- * legitimately-empty middle cells. Only the leading and trailing empty strings
- * produced by splitting a `| a | b |`-style row are stripped — a prior
- * `.filter(c => c.length > 0)` approach also dropped empty interior cells,
- * which silently collapsed the column count and caused `cols.length >= N`
- * guards to misfire (e.g. a voice-dump row with an empty tiers cell).
- */
-function parseTableRow(row: string): string[] {
-  const parts = row.split("|");
-  // Strip the leading empty produced by a leading `|`, if present
-  if (parts.length > 0 && parts[0].trim() === "") parts.shift();
-  // Strip the trailing empty produced by a trailing `|`, if present
-  if (parts.length > 0 && parts[parts.length - 1].trim() === "") parts.pop();
-  return parts.map((c) => c.trim());
-}
-
 // ──────────────────────────────────────────────────────────────────────────────
 // BM25 Index Cache (per-request, lazy)
 // ──────────────────────────────────────────────────────────────────────────────
diff --git a/workers/test/governance-parser.test.mjs b/workers/test/governance-parser.test.mjs
index 4ae40c5..a3cc414 100644
--- a/workers/test/governance-parser.test.mjs
+++ b/workers/test/governance-parser.test.mjs
@@ -36,6 +36,7 @@ const ARTICLE_PATHS = {
   basePrerequisites: "odd/challenge/base-prerequisites.md",
   normativeVocabulary: "odd/challenge/normative-vocabulary.md",
   stakesCalibration: "odd/challenge/stakes-calibration.md",
+  telemetryGovernance: "canon/constraints/telemetry-governance.md",
 };
 
 async function fetchArticle(path) {
@@ -339,6 +340,47 @@ async function run() {
   ok("planning has baseline+elevated", calib.get("planning")?.tiers.length === 2);
   ok("execution has all three tiers", calib.get("execution")?.tiers.length === 3);
 
+  console.log("\n─── Test 8: Self-report headers table (telemetry_policy canary) ───");
+  // Mirrors parseSelfReportHeadersTable in workers/src/index.ts. If either
+  // parser changes, both must change — tracked as a known duplication per
+  // PR #106 discussion.
+  const parseHeaders = (md) => {
+    const section = md.match(/###\s+Self-Report Fields[^\n]*\n([\s\S]*?)(?=\n###|\n##|$)/);
+    if (!section) return null;
+    const out = {};
+    for (const raw of section[1].split("\n")) {
+      if (!raw.includes("|")) continue;
+      const parts = raw.split("|");
+      if (parts.length > 0 && parts[0].trim() === "") parts.shift();
+      if (parts.length > 0 && parts[parts.length - 1].trim() === "") parts.pop();
+      const cols = parts.map((c) => c.trim());
+      if (cols.length < 4) continue;
+      const headerName = cols[1].replace(/`/g, "").trim();
+      if (!headerName.startsWith("x-oddkit-")) continue;
+      const description = cols[3].trim();
+      if (!description) continue;
+      out[headerName] = description;
+    }
+    return Object.keys(out).length > 0 ? out : null;
+  };
+
+  const headers = parseHeaders(articles.telemetryGovernance);
+  ok("self-report headers parse", headers !== null);
+  ok("eight headers extracted", headers && Object.keys(headers).length === 8, `got ${headers ? Object.keys(headers).length : 0}`);
+  ok("x-oddkit-client present", headers && typeof headers["x-oddkit-client"] === "string" && headers["x-oddkit-client"].length > 0);
+  ok("x-oddkit-surface present", headers && typeof headers["x-oddkit-surface"] === "string" && headers["x-oddkit-surface"].length > 0);
+  ok("x-oddkit-capabilities present", headers && typeof headers["x-oddkit-capabilities"] === "string" && headers["x-oddkit-capabilities"].length > 0);
+  ok(
+    "descriptions are non-trivial (canon Description column, not Field label)",
+    headers && Object.values(headers).every((d) => d.length > 15),
+    `shortest: ${headers ? Math.min(...Object.values(headers).map((d) => d.length)) : 0} chars`,
+  );
+
+  // Degradation: missing section → null
+  ok("no section returns null", parseHeaders("# No section here\n") === null);
+  // Degradation: section present but no table rows → null
+  ok("empty section returns null", parseHeaders("### Self-Report Fields\n\n(no table)\n") === null);
+
   console.log(`\n${passed} passed, ${failed} failed`);
   process.exit(failed === 0 ? 0 : 1);
 }