diff --git a/.github/workflows/actions-fleet-monitor.yml b/.github/workflows/actions-fleet-monitor.yml
index f7f44b83e..81f309570 100644
--- a/.github/workflows/actions-fleet-monitor.yml
+++ b/.github/workflows/actions-fleet-monitor.yml
@@ -80,11 +80,11 @@ jobs:
         with:
           repository: petry-projects/.github-private
           ref: main  # Always checkout .github-private@main; workflow_call callers target @main per docs
-          token: ${{ secrets.GH_PAT_WORKFLOWS }}
+          token: ${{ secrets.GH_PAT_DON_PETRY || secrets.GH_PAT_WORKFLOWS }}
 
       - name: Run fleet monitor
         env:
-          GH_TOKEN: ${{ secrets.GH_PAT_WORKFLOWS }}
+          GH_TOKEN: ${{ secrets.GH_PAT_DON_PETRY || secrets.GH_PAT_WORKFLOWS }}
         run: bash scripts/fleet_monitor.sh
 
       - name: Summarize token usage by workflow
@@ -94,7 +94,7 @@ jobs:
         # tracking + the fleet-tracker auto-close). Let it fail soft.
         continue-on-error: true
         env:
-          GH_TOKEN: ${{ secrets.GH_PAT_WORKFLOWS }}
+          GH_TOKEN: ${{ secrets.GH_PAT_DON_PETRY || secrets.GH_PAT_WORKFLOWS }}
           REPO: petry-projects/.github-private
           LOOKBACK_DAYS: ${{ env.LOOKBACK_DAYS }}
         run: |
@@ -226,7 +226,7 @@ jobs:
       - name: Migrate misaligned fleet-tracker issues
         uses: actions/github-script@3a2844b7e9c422d3c10d287c895573f7108da1b3  # v9.0.0
         env:
-          GH_PAT_WORKFLOWS: ${{ secrets.GH_PAT_WORKFLOWS }}
+          GH_PAT_WORKFLOWS: ${{ secrets.GH_PAT_DON_PETRY || secrets.GH_PAT_WORKFLOWS }}
         with:
           github-token: ${{ env.GH_PAT_WORKFLOWS }}
           script: |
@@ -286,7 +286,7 @@ jobs:
         uses: actions/github-script@3a2844b7e9c422d3c10d287c895573f7108da1b3  # v9.0.0
         env:
           ORG: ${{ env.ORG }}
-          GH_PAT_WORKFLOWS: ${{ secrets.GH_PAT_WORKFLOWS }}
+          GH_PAT_WORKFLOWS: ${{ secrets.GH_PAT_DON_PETRY || secrets.GH_PAT_WORKFLOWS }}
         with:
           # Use GH_PAT_WORKFLOWS for cross-repo issues:write on target repos
           github-token: ${{ env.GH_PAT_WORKFLOWS }}
@@ -409,7 +409,7 @@ jobs:
         env:
           ORG: ${{ env.ORG }}
           STALE_DAYS: '3'
-          GH_PAT_WORKFLOWS: ${{ secrets.GH_PAT_WORKFLOWS }}
+          GH_PAT_WORKFLOWS: ${{ secrets.GH_PAT_DON_PETRY || secrets.GH_PAT_WORKFLOWS }}
         with:
           # Use GH_PAT_WORKFLOWS for cross-repo issues:write on target repos
           github-token: ${{ env.GH_PAT_WORKFLOWS }}
@@ -467,7 +467,7 @@ jobs:
       - name: Dispatch dev-lead for critical items
         if: env.HIGH_FAILURE_COUNT != '' && env.HIGH_FAILURE_COUNT != '0'
         env:
-          GH_TOKEN: ${{ secrets.GH_PAT_WORKFLOWS }}
+          GH_TOKEN: ${{ secrets.GH_PAT_DON_PETRY || secrets.GH_PAT_WORKFLOWS }}
         run: |
           # github.token events do not trigger issues:labeled workflows; use the
           # PAT to fire repository_dispatch so Dev-Lead Agent is explicitly woken.
@@ -556,7 +556,7 @@ jobs:
       - name: Dispatch dev-lead for stub drift
         if: env.HAS_STUB_DRIFT == 'true'
         env:
-          GH_TOKEN: ${{ secrets.GH_PAT_WORKFLOWS }}
+          GH_TOKEN: ${{ secrets.GH_PAT_DON_PETRY || secrets.GH_PAT_WORKFLOWS }}
         run: |
           # github.token events do not trigger issues:labeled workflows; use the
           # PAT to fire repository_dispatch so Dev-Lead Agent is explicitly woken.
@@ -585,7 +585,7 @@ jobs:
         env:
           EVENT_NAME: ${{ github.event_name }}
           REMEDIATE_MODE: ${{ inputs.remediate || 'dry-run' }}
-          REMEDIATE_TOKEN: ${{ secrets.GH_PAT_WORKFLOWS }}
+          REMEDIATE_TOKEN: ${{ secrets.GH_PAT_DON_PETRY || secrets.GH_PAT_WORKFLOWS }}
           REMEDIATE_PILOT_REPO: ${{ vars.FLEET_REMEDIATE_PILOT_REPO }}
           REMEDIATE_MAX_REPOS: ${{ vars.FLEET_REMEDIATE_MAX_REPOS || '1' }}
         run: |
@@ -617,7 +617,7 @@ jobs:
         # continue-on-error keeps this informational step non-blocking.
         continue-on-error: true
         env:
-          GH_TOKEN: ${{ secrets.GH_PAT_WORKFLOWS }}
+          GH_TOKEN: ${{ secrets.GH_PAT_DON_PETRY || secrets.GH_PAT_WORKFLOWS }}
           ORG: ${{ env.ORG }}
           LOOKBACK_DAYS: ${{ env.LOOKBACK_DAYS }}
         run: bash scripts/token_report.sh >> "$GITHUB_STEP_SUMMARY"
diff --git a/.github/workflows/auto-rebase-retry.yml b/.github/workflows/auto-rebase-retry.yml
index 1e19737f2..2d7c56137 100644
--- a/.github/workflows/auto-rebase-retry.yml
+++ b/.github/workflows/auto-rebase-retry.yml
@@ -46,7 +46,7 @@ jobs:
           # workflow_run events (GitHub's recursion guard), so only GH_PAT_WORKFLOWS
           # lets attempts past the first chain back into this handler toward
           # MAX_ATTEMPTS. With github.token alone, a single retry still fires.
-          GH_TOKEN: ${{ secrets.GH_PAT_WORKFLOWS || github.token }}
+          GH_TOKEN: ${{ secrets.GH_PAT_DON_PETRY || secrets.GH_PAT_WORKFLOWS || github.token }}
           REPO: ${{ github.repository }}
           CONCLUSION: ${{ github.event.workflow_run.conclusion }}
           RUN_ID: ${{ github.event.workflow_run.id }}
diff --git a/.github/workflows/dependency-advisory.yml b/.github/workflows/dependency-advisory.yml
index e7a30d507..2ace05a9d 100644
--- a/.github/workflows/dependency-advisory.yml
+++ b/.github/workflows/dependency-advisory.yml
@@ -45,7 +45,8 @@ jobs:
       - name: Checkout
         uses: actions/checkout@df4cb1c069e1874edd31b4311f1884172cec0e10  # v6.0.3
         with:
-          token: ${{ secrets.GH_PAT_WORKFLOWS || secrets.GITHUB_TOKEN }}
+          token: ${{ secrets.GH_PAT_DON_PETRY || secrets.GH_PAT_WORKFLOWS || secrets.GITHUB_TOKEN }}
+          persist-credentials: false
 
       - name: Set up Node
         uses: actions/setup-node@48b55a011bda9f5d6aeb4c2d9c7362e8dae4041e  # v6.4.0
@@ -72,5 +73,5 @@ jobs:
 
       - name: Run dependency advisory
         env:
-          GH_TOKEN: ${{ secrets.GH_PAT_WORKFLOWS || secrets.GITHUB_TOKEN }}
+          GH_TOKEN: ${{ secrets.GH_PAT_DON_PETRY || secrets.GH_PAT_WORKFLOWS || secrets.GITHUB_TOKEN }}
         run: bash scripts/aw-dependency-advisory.sh
diff --git a/.github/workflows/dev-lead-retry.yml b/.github/workflows/dev-lead-retry.yml
index e6439648e..34126cd45 100644
--- a/.github/workflows/dev-lead-retry.yml
+++ b/.github/workflows/dev-lead-retry.yml
@@ -35,7 +35,7 @@ concurrency:
   cancel-in-progress: false
 
 env:
-  GH_TOKEN: ${{ secrets.GH_PAT_WORKFLOWS || secrets.GITHUB_TOKEN }}
+  GH_TOKEN: ${{ secrets.GH_PAT_DON_PETRY || secrets.GH_PAT_WORKFLOWS || secrets.GITHUB_TOKEN }}
   TARGET_ORG: ${{ vars.TARGET_ORG || 'petry-projects' }}
   DRY_RUN: ${{ inputs.dry_run || 'false' }}
   DISPATCH_DELAY_SEC: ${{ inputs.dispatch_delay_sec || '30' }}
@@ -48,7 +48,8 @@ jobs:
       - name: Checkout agent repo
         uses: actions/checkout@df4cb1c069e1874edd31b4311f1884172cec0e10 # v6.0.3
         with:
-          token: ${{ secrets.GH_PAT_WORKFLOWS || secrets.GITHUB_TOKEN }}
+          token: ${{ secrets.GH_PAT_DON_PETRY || secrets.GH_PAT_WORKFLOWS || secrets.GITHUB_TOKEN }}
+          persist-credentials: false
 
       - name: Scan and retry rate-limited PRs
         run: bash scripts/dev-lead-retry.sh
diff --git a/.github/workflows/dev-lead-reusable.yml b/.github/workflows/dev-lead-reusable.yml
index db57e330a..5cde5faba 100644
--- a/.github/workflows/dev-lead-reusable.yml
+++ b/.github/workflows/dev-lead-reusable.yml
@@ -20,14 +20,27 @@ on:
     secrets:
       CLAUDE_CODE_OAUTH_TOKEN:
         required: true
+      GH_PAT_DON_PETRY:
+        # don-petry's write/workflows PAT (the account dev-lead commits/pushes as).
+        # Canonical name under the GH_PAT_<ACCOUNT> schema; supersedes GH_PAT_WORKFLOWS,
+        # which is kept below as a transition fallback.
+        required: false
       GH_PAT_WORKFLOWS:
+        # DEPRECATED alias of GH_PAT_DON_PETRY — kept during the rename migration so
+        # callers that haven't set the new secret yet still authenticate. Remove once
+        # GH_PAT_DON_PETRY is confirmed everywhere.
         # Required for cross-repo callers: the default github.token cannot read
         # petry-projects/.github-private from another repo. Callers must pass a
         # PAT that has read access to petry-projects/.github-private.
         required: false
       GOOGLE_API_KEY:
         required: false
+      GH_PAT_DON_PETRY_COPILOT:
+        # don-petry's Copilot-subscription PAT (Copilot review engine). Canonical
+        # name; supersedes GH_PAT, kept below as a transition fallback.
+        required: false
       GH_PAT:
+        # DEPRECATED alias of GH_PAT_DON_PETRY_COPILOT — kept during migration.
         required: false
 
 permissions: {}
@@ -98,14 +111,21 @@ jobs:
       checks: read
       statuses: read
     env:
-      BOT_USER: ${{ vars.BOT_USER || 'donpetry-bot' }}
+      # dev-lead's acting identity is DECLARED in personas/dev-lead/persona.yml
+      # (runtime.identity.account) and resolved into BOT_USER by the "Resolve
+      # persona identity" step below. This job-level value is the hardcoded
+      # fail-safe default (don-petry) used only if that step cannot read the
+      # manifest. It does NOT consult vars.BOT_USER — that shared variable is
+      # what caused the donpetry-bot regression (#1316); the manifest is now the
+      # sole source of truth. See scripts/lib/resolve-persona-identity.sh.
+      BOT_USER: 'don-petry'
       TRUSTED_BOTS: ${{ vars.TRUSTED_BOTS || 'copilot-pull-request-reviewer[bot],gemini-code-assist[bot],sonarqubecloud[bot],coderabbitai[bot],chatgpt-codex-connector[bot]' }}
       TRIGGER_PHRASES: ${{ vars.TRIGGER_PHRASES || '@dev-lead' }}
       DEV_LEAD_ENGINE: ${{ vars.DEV_LEAD_ENGINE || 'claude' }}
       DEV_LEAD_DRY_RUN: ${{ vars.DEV_LEAD_DRY_RUN || 'false' }}
       DEEP_TIMEOUT_SEC: ${{ vars.DEEP_TIMEOUT_SEC || '' }}
       ACTION_TIMEOUT_SEC: ${{ vars.ACTION_TIMEOUT_SEC || '' }}
-      GH_TOKEN: ${{ secrets.GH_PAT_WORKFLOWS || github.token }}
+      GH_TOKEN: ${{ secrets.GH_PAT_DON_PETRY || secrets.GH_PAT_WORKFLOWS || github.token }}
       GEMINI_CLI_TRUST_WORKSPACE: "true"
       DEV_LEAD_SCRIPTS: .dev-lead/scripts
       DEV_LEAD_PROMPTS: .dev-lead/prompts/dev-lead
@@ -116,7 +136,7 @@ jobs:
       - name: Checkout caller repo
         uses: actions/checkout@df4cb1c069e1874edd31b4311f1884172cec0e10 # v6.0.3
         with:
-          token: ${{ secrets.GH_PAT_WORKFLOWS || github.token }}
+          token: ${{ secrets.GH_PAT_DON_PETRY || secrets.GH_PAT_WORKFLOWS || github.token }}
 
       - name: Checkout dev-lead scripts
         uses: actions/checkout@df4cb1c069e1874edd31b4311f1884172cec0e10 # v6.0.3
@@ -128,10 +148,11 @@ jobs:
           path: .dev-lead
           # Cross-repo callers: github.token cannot read a private org repo — ensure
           # GH_PAT_WORKFLOWS is set as an org or repo secret in the calling repo.
-          token: ${{ secrets.GH_PAT_WORKFLOWS || github.token }}
+          token: ${{ secrets.GH_PAT_DON_PETRY || secrets.GH_PAT_WORKFLOWS || github.token }}
           sparse-checkout: |
             scripts
             prompts/dev-lead
+            personas/dev-lead
 
       - name: Protect dev-lead scripts from accidental git staging
         run: echo ".dev-lead/" >> .gitignore
@@ -141,6 +162,20 @@ jobs:
           git checkout .gitignore || true
           echo ".dev-lead/" >> .git/info/exclude
 
+      - name: Resolve persona identity
+        # Source of truth for who dev-lead acts as: personas/dev-lead/persona.yml
+        # (runtime.identity.account). Exporting BOT_USER here overrides the job-env
+        # fallback for every later step — intent classification, git-identity, and
+        # the commit/push. On resolve failure we keep the job default (don-petry)
+        # and warn, rather than silently reverting to the shared machine user.
+        run: |
+          if account="$(bash .dev-lead/scripts/lib/resolve-persona-identity.sh dev-lead .dev-lead/personas account)"; then
+            echo "BOT_USER=$account" >> "$GITHUB_ENV"
+            echo "resolved dev-lead identity: BOT_USER=$account"
+          else
+            echo "::warning::could not resolve dev-lead identity from manifest; using fallback BOT_USER=$BOT_USER"
+          fi
+
       - name: Classify event intent
         id: intent
         env:
@@ -148,7 +183,7 @@ jobs:
           CLAUDE_CODE_OAUTH_TOKEN: ${{ secrets.CLAUDE_CODE_OAUTH_TOKEN }}
           GOOGLE_API_KEY: ${{ secrets.GOOGLE_API_KEY }}
           GEMINI_API_KEY: ${{ secrets.GOOGLE_API_KEY }}
-          COPILOT_GITHUB_TOKEN: ${{ secrets.GH_PAT }}
+          COPILOT_GITHUB_TOKEN: ${{ secrets.GH_PAT_DON_PETRY_COPILOT || secrets.GH_PAT }}
         run: bash .dev-lead/scripts/dev-lead-intent.sh
 
       - name: Log intent
@@ -185,10 +220,10 @@ jobs:
         if: env.INTENT_TYPE != 'skip' && env.INTENT_TYPE != 'enable-auto-merge'
         env:
           CLAUDE_CODE_OAUTH_TOKEN: ${{ secrets.CLAUDE_CODE_OAUTH_TOKEN }}
-          GH_PAT_WORKFLOWS: ${{ secrets.GH_PAT_WORKFLOWS }}
+          GH_PAT_WORKFLOWS: ${{ secrets.GH_PAT_DON_PETRY || secrets.GH_PAT_WORKFLOWS }}
           GOOGLE_API_KEY: ${{ secrets.GOOGLE_API_KEY }}
           GEMINI_API_KEY: ${{ secrets.GOOGLE_API_KEY }}
-          GH_PAT: ${{ secrets.GH_PAT }}
+          GH_PAT: ${{ secrets.GH_PAT_DON_PETRY_COPILOT || secrets.GH_PAT }}
         run: bash .dev-lead/scripts/dev-lead-preflight.sh
 
       # ── Install engine CLIs ───────────────────────────────────────────────
@@ -244,7 +279,7 @@ jobs:
           CLAUDE_CODE_OAUTH_TOKEN: ${{ secrets.CLAUDE_CODE_OAUTH_TOKEN }}
           GOOGLE_API_KEY: ${{ secrets.GOOGLE_API_KEY }}
           GEMINI_API_KEY: ${{ secrets.GOOGLE_API_KEY }}
-          COPILOT_GITHUB_TOKEN: ${{ secrets.GH_PAT }}
+          COPILOT_GITHUB_TOKEN: ${{ secrets.GH_PAT_DON_PETRY_COPILOT || secrets.GH_PAT }}
           PROMPTS_DIR: .dev-lead/prompts/dev-lead
         run: |
           bash .dev-lead/scripts/dev-lead-fix-ci.sh || rc=$?
@@ -263,7 +298,7 @@ jobs:
           CLAUDE_CODE_OAUTH_TOKEN: ${{ secrets.CLAUDE_CODE_OAUTH_TOKEN }}
           GOOGLE_API_KEY: ${{ secrets.GOOGLE_API_KEY }}
           GEMINI_API_KEY: ${{ secrets.GOOGLE_API_KEY }}
-          COPILOT_GITHUB_TOKEN: ${{ secrets.GH_PAT }}
+          COPILOT_GITHUB_TOKEN: ${{ secrets.GH_PAT_DON_PETRY_COPILOT || secrets.GH_PAT }}
           PROMPTS_DIR: .dev-lead/prompts/dev-lead
         run: |
           bash .dev-lead/scripts/dev-lead-fix-reviews.sh || rc=$?
@@ -282,7 +317,7 @@ jobs:
           CLAUDE_CODE_OAUTH_TOKEN: ${{ secrets.CLAUDE_CODE_OAUTH_TOKEN }}
           GOOGLE_API_KEY: ${{ secrets.GOOGLE_API_KEY }}
           GEMINI_API_KEY: ${{ secrets.GOOGLE_API_KEY }}
-          COPILOT_GITHUB_TOKEN: ${{ secrets.GH_PAT }}
+          COPILOT_GITHUB_TOKEN: ${{ secrets.GH_PAT_DON_PETRY_COPILOT || secrets.GH_PAT }}
           PROMPTS_DIR: .dev-lead/prompts/dev-lead
         run: |
           bash .dev-lead/scripts/dev-lead-fix-reviews.sh || rc=$?
@@ -301,7 +336,7 @@ jobs:
           CLAUDE_CODE_OAUTH_TOKEN: ${{ secrets.CLAUDE_CODE_OAUTH_TOKEN }}
           GOOGLE_API_KEY: ${{ secrets.GOOGLE_API_KEY }}
           GEMINI_API_KEY: ${{ secrets.GOOGLE_API_KEY }}
-          COPILOT_GITHUB_TOKEN: ${{ secrets.GH_PAT }}
+          COPILOT_GITHUB_TOKEN: ${{ secrets.GH_PAT_DON_PETRY_COPILOT || secrets.GH_PAT }}
           PROMPTS_DIR: .dev-lead/prompts/dev-lead
         run: |
           bash .dev-lead/scripts/dev-lead-fix-reviews.sh || rc=$?
@@ -320,7 +355,7 @@ jobs:
           CLAUDE_CODE_OAUTH_TOKEN: ${{ secrets.CLAUDE_CODE_OAUTH_TOKEN }}
           GOOGLE_API_KEY: ${{ secrets.GOOGLE_API_KEY }}
           GEMINI_API_KEY: ${{ secrets.GOOGLE_API_KEY }}
-          COPILOT_GITHUB_TOKEN: ${{ secrets.GH_PAT }}
+          COPILOT_GITHUB_TOKEN: ${{ secrets.GH_PAT_DON_PETRY_COPILOT || secrets.GH_PAT }}
           PROMPTS_DIR: .dev-lead/prompts/dev-lead
         run: |
           bash .dev-lead/scripts/dev-lead-fix-reviews.sh || rc=$?
@@ -337,7 +372,7 @@ jobs:
           CLAUDE_CODE_OAUTH_TOKEN: ${{ secrets.CLAUDE_CODE_OAUTH_TOKEN }}
           GOOGLE_API_KEY: ${{ secrets.GOOGLE_API_KEY }}
           GEMINI_API_KEY: ${{ secrets.GOOGLE_API_KEY }}
-          COPILOT_GITHUB_TOKEN: ${{ secrets.GH_PAT }}
+          COPILOT_GITHUB_TOKEN: ${{ secrets.GH_PAT_DON_PETRY_COPILOT || secrets.GH_PAT }}
           PROMPTS_DIR: .dev-lead/prompts/dev-lead
         run: |
           bash .dev-lead/scripts/dev-lead-fix-issue.sh || rc=$?
@@ -354,7 +389,7 @@ jobs:
           CLAUDE_CODE_OAUTH_TOKEN: ${{ secrets.CLAUDE_CODE_OAUTH_TOKEN }}
           GOOGLE_API_KEY: ${{ secrets.GOOGLE_API_KEY }}
           GEMINI_API_KEY: ${{ secrets.GOOGLE_API_KEY }}
-          COPILOT_GITHUB_TOKEN: ${{ secrets.GH_PAT }}
+          COPILOT_GITHUB_TOKEN: ${{ secrets.GH_PAT_DON_PETRY_COPILOT || secrets.GH_PAT }}
           PROMPTS_DIR: .dev-lead/prompts/dev-lead
         run: |
           bash .dev-lead/scripts/dev-lead-fix-reviews.sh || rc=$?
@@ -400,7 +435,7 @@ jobs:
       - name: Check for associated non-fork PR
         id: check-pr
         env:
-          GH_TOKEN: ${{ secrets.GH_PAT_WORKFLOWS || github.token }}
+          GH_TOKEN: ${{ secrets.GH_PAT_DON_PETRY || secrets.GH_PAT_WORKFLOWS || github.token }}
           CHECK_RUN_PRS: ${{ toJson(github.event.check_run.pull_requests) }}
           CHECK_RUN_HEAD_SHA: ${{ github.event.check_run.head_sha }}
           REPO_FULL_NAME: ${{ github.repository }}
@@ -446,7 +481,7 @@ jobs:
       - name: Emit repository_dispatch dev-lead-ci-failure
         if: steps.check-pr.outputs.should_relay == 'true'
         env:
-          GH_TOKEN: ${{ secrets.GH_PAT_WORKFLOWS || github.token }}
+          GH_TOKEN: ${{ secrets.GH_PAT_DON_PETRY || secrets.GH_PAT_WORKFLOWS || github.token }}
           RELAY_PR_NUMBER: ${{ steps.check-pr.outputs.pr_number }}
           RELAY_HEAD_SHA: ${{ github.event.check_run.head_sha }}
           RELAY_REPO: ${{ github.repository }}
diff --git a/.github/workflows/feature-ideation.yml b/.github/workflows/feature-ideation.yml
index 9263fd1f1..b68d48af4 100644
--- a/.github/workflows/feature-ideation.yml
+++ b/.github/workflows/feature-ideation.yml
@@ -125,7 +125,7 @@ jobs:
         # NEVER starts a run (loop prevention) — the enhancement would silently
         # no-op. Same constraint as scripts/initiative-planner/redispatch.sh.
         env:
-          GH_TOKEN: ${{ secrets.GH_PAT_WORKFLOWS }}
+          GH_TOKEN: ${{ secrets.GH_PAT_DON_PETRY || secrets.GH_PAT_WORKFLOWS }}
         run: |
           if [ -z "${GH_TOKEN}" ]; then
             echo "::error::GH_PAT_WORKFLOWS is required — a workflow_dispatch fired with GITHUB_TOKEN will not start a run."
@@ -134,7 +134,7 @@ jobs:
 
       - name: Re-dispatch under workflow_dispatch
         env:
-          GH_TOKEN: ${{ secrets.GH_PAT_WORKFLOWS }}
+          GH_TOKEN: ${{ secrets.GH_PAT_DON_PETRY || secrets.GH_PAT_WORKFLOWS }}
           REPO: ${{ github.repository }}
           # Discussion numbers are always positive integers, so this is safe to
           # forward directly to the dispatch input.
diff --git a/.github/workflows/gh-aw-cross-org.yml b/.github/workflows/gh-aw-cross-org.yml
index e01799733..a052db53a 100644
--- a/.github/workflows/gh-aw-cross-org.yml
+++ b/.github/workflows/gh-aw-cross-org.yml
@@ -34,7 +34,7 @@ env:
   BOT_USER: ${{ vars.BOT_USER || 'donpetry-bot' }}
   DEV_LEAD_ENGINE: ${{ vars.DEV_LEAD_ENGINE || 'claude' }}
   DEV_LEAD_DRY_RUN: ${{ vars.DEV_LEAD_DRY_RUN || 'false' }}
-  GH_TOKEN: ${{ secrets.GH_PAT_WORKFLOWS }}
+  GH_TOKEN: ${{ secrets.GH_PAT_DON_PETRY || secrets.GH_PAT_WORKFLOWS }}
 
 jobs:
   dispatch:
@@ -45,7 +45,7 @@ jobs:
       GOOGLE_API_KEY: ${{ secrets.GOOGLE_API_KEY }}
       GEMINI_API_KEY: ${{ secrets.GOOGLE_API_KEY }}
       GEMINI_CLI_TRUST_WORKSPACE: "true"
-      COPILOT_GITHUB_TOKEN: ${{ secrets.GH_PAT }}
+      COPILOT_GITHUB_TOKEN: ${{ secrets.GH_PAT_DON_PETRY_COPILOT || secrets.GH_PAT }}
 
     steps:
       - name: Validate source repo
@@ -62,7 +62,7 @@ jobs:
       - name: Checkout .github-private (scripts and prompts)
         uses: actions/checkout@df4cb1c069e1874edd31b4311f1884172cec0e10 # v6.0.3
         with:
-          token: ${{ secrets.GH_PAT_WORKFLOWS || secrets.GITHUB_TOKEN }}
+          token: ${{ secrets.GH_PAT_DON_PETRY || secrets.GH_PAT_WORKFLOWS || secrets.GITHUB_TOKEN }}
           path: dotgithub-private
 
       # Checkout the target (source) repo so the engine can read and modify its code.
@@ -71,7 +71,7 @@ jobs:
         uses: actions/checkout@df4cb1c069e1874edd31b4311f1884172cec0e10 # v6.0.3
         with:
           repository: ${{ github.event.client_payload.source_repo }}
-          token: ${{ secrets.GH_PAT_WORKFLOWS }}
+          token: ${{ secrets.GH_PAT_DON_PETRY || secrets.GH_PAT_WORKFLOWS }}
           path: target-repo
 
       - name: Configure git identity for target repo
@@ -89,7 +89,7 @@ jobs:
         env:
           CLAUDE_CODE_VERSION: ${{ vars.CLAUDE_CODE_VERSION || 'latest' }}
           GOOGLE_API_KEY: ${{ secrets.GOOGLE_API_KEY }}
-          GH_TOKEN: ${{ secrets.GH_PAT || secrets.GITHUB_TOKEN }}
+          GH_TOKEN: ${{ secrets.GH_PAT_DON_PETRY_COPILOT || secrets.GH_PAT || secrets.GITHUB_TOKEN }}
         run: |
           set -euo pipefail
           mkdir -p "$HOME/.npm-global"
@@ -127,7 +127,7 @@ jobs:
           REPO: ${{ github.event.client_payload.source_repo }}
           PROMPTS_DIR: ${{ github.workspace }}/dotgithub-private/prompts/dev-lead
           REVIEW_ENGINE: ${{ env.DEV_LEAD_ENGINE }}
-          GH_TOKEN: ${{ secrets.GH_PAT_WORKFLOWS }}
+          GH_TOKEN: ${{ secrets.GH_PAT_DON_PETRY || secrets.GH_PAT_WORKFLOWS }}
         working-directory: target-repo
         run: bash ${{ github.workspace }}/dotgithub-private/scripts/dev-lead-fix-issue.sh
 
@@ -144,6 +144,6 @@ jobs:
           REPO: ${{ github.event.client_payload.source_repo }}
           PROMPTS_DIR: ${{ github.workspace }}/dotgithub-private/prompts/dev-lead
           REVIEW_ENGINE: ${{ env.DEV_LEAD_ENGINE }}
-          GH_TOKEN: ${{ secrets.GH_PAT_WORKFLOWS }}
+          GH_TOKEN: ${{ secrets.GH_PAT_DON_PETRY || secrets.GH_PAT_WORKFLOWS }}
         working-directory: target-repo
         run: bash ${{ github.workspace }}/dotgithub-private/scripts/dev-lead-fix-ci.sh
diff --git a/.github/workflows/idea-triage.yml b/.github/workflows/idea-triage.yml
index cc62853c6..ca3fe7b26 100644
--- a/.github/workflows/idea-triage.yml
+++ b/.github/workflows/idea-triage.yml
@@ -65,7 +65,7 @@ jobs:
       # the self path only; cross-repo runs use GH_PAT_WORKFLOWS (same pattern as
       # initiative-planner.yml). S1 (#818) will replace this with a minted GitHub
       # App token for tighter scoping.
-      TARGET_GH_TOKEN: ${{ (github.event.inputs.target_repo == '' || github.event.inputs.target_repo == github.repository) && secrets.GITHUB_TOKEN || secrets.GH_PAT_WORKFLOWS }}
+      TARGET_GH_TOKEN: ${{ (github.event.inputs.target_repo == '' || github.event.inputs.target_repo == github.repository) && secrets.GITHUB_TOKEN || secrets.GH_PAT_DON_PETRY || secrets.GH_PAT_WORKFLOWS }}
       DRY_RUN: ${{ inputs.dry_run && '1' || '0' }}
     steps:
       - name: Checkout agent repo
diff --git a/.github/workflows/initiative-driver-canary.yml b/.github/workflows/initiative-driver-canary.yml
index 60f07f914..bf04b57a2 100644
--- a/.github/workflows/initiative-driver-canary.yml
+++ b/.github/workflows/initiative-driver-canary.yml
@@ -81,7 +81,7 @@ jobs:
         if: steps.guard.outputs.configured == 'true'
         env:
           # PAT required — a workflow_dispatch fired with GITHUB_TOKEN never starts a run.
-          GH_TOKEN: ${{ secrets.GH_PAT_WORKFLOWS }}
+          GH_TOKEN: ${{ secrets.GH_PAT_DON_PETRY || secrets.GH_PAT_WORKFLOWS }}
           REPO: ${{ github.repository }}
           CANARY_OUT: initiative_driver_canary.md
         run: bash scripts/initiative_driver_canary.sh
diff --git a/.github/workflows/initiative-driver.yml b/.github/workflows/initiative-driver.yml
index b9d85f06a..0b1641263 100644
--- a/.github/workflows/initiative-driver.yml
+++ b/.github/workflows/initiative-driver.yml
@@ -88,7 +88,7 @@ jobs:
       - name: Guard — PAT present
         # PAT required: a label applied with GITHUB_TOKEN does NOT trigger dev-lead.yml.
         env:
-          GH_TOKEN: ${{ secrets.GH_PAT_WORKFLOWS }}
+          GH_TOKEN: ${{ secrets.GH_PAT_DON_PETRY || secrets.GH_PAT_WORKFLOWS }}
         run: |
           if [ -z "${GH_TOKEN}" ]; then
             echo "::error::GH_PAT_WORKFLOWS is required — a label applied with GITHUB_TOKEN will not trigger dev-lead."
@@ -97,5 +97,5 @@ jobs:
 
       - name: Drive initiative
         env:
-          GH_TOKEN: ${{ secrets.GH_PAT_WORKFLOWS }}
+          GH_TOKEN: ${{ secrets.GH_PAT_DON_PETRY || secrets.GH_PAT_WORKFLOWS }}
         run: bash scripts/initiative-driver.sh
diff --git a/.github/workflows/initiative-planner-canary.yml b/.github/workflows/initiative-planner-canary.yml
index c80e0049d..eb1191a01 100644
--- a/.github/workflows/initiative-planner-canary.yml
+++ b/.github/workflows/initiative-planner-canary.yml
@@ -67,7 +67,7 @@ jobs:
         if: steps.guard.outputs.configured == 'true'
         env:
           # PAT required — a workflow_dispatch fired with GITHUB_TOKEN never starts a run.
-          GH_TOKEN: ${{ secrets.GH_PAT_WORKFLOWS }}
+          GH_TOKEN: ${{ secrets.GH_PAT_DON_PETRY || secrets.GH_PAT_WORKFLOWS }}
           REPO: ${{ github.repository }}
           CANARY_OUT: initiative_canary.md
         run: bash scripts/initiative_canary.sh
diff --git a/.github/workflows/initiative-planner.yml b/.github/workflows/initiative-planner.yml
index 918fcba66..36fb0a1e9 100644
--- a/.github/workflows/initiative-planner.yml
+++ b/.github/workflows/initiative-planner.yml
@@ -99,7 +99,7 @@ jobs:
         # PAT required: a workflow_dispatch fired with GITHUB_TOKEN is accepted but
         # NEVER starts a run (loop prevention) — the planner would silently no-op.
         env:
-          GH_TOKEN: ${{ secrets.GH_PAT_WORKFLOWS }}
+          GH_TOKEN: ${{ secrets.GH_PAT_DON_PETRY || secrets.GH_PAT_WORKFLOWS }}
         run: |
           if [ -z "${GH_TOKEN}" ]; then
             echo "::error::GH_PAT_WORKFLOWS is required — a workflow_dispatch fired with GITHUB_TOKEN will not start a run."
@@ -108,7 +108,7 @@ jobs:
 
       - name: Re-dispatch under workflow_dispatch
         env:
-          GH_TOKEN: ${{ secrets.GH_PAT_WORKFLOWS }}
+          GH_TOKEN: ${{ secrets.GH_PAT_DON_PETRY || secrets.GH_PAT_WORKFLOWS }}
           REPO: ${{ github.repository }}
           # Dogfood path: this repo plans its own discussions, so the target is
           # self. Fleet callers (#820) live in their own repos and set their host.
@@ -144,7 +144,7 @@ jobs:
       # S1 (#818) will replace this with a minted GitHub App token for tighter scoping.
       # The workflow_dispatch re-invoke keeps GH_PAT_WORKFLOWS (redispatch job) — that
       # role is loop-prevention, not content writes, and is unaffected.
-      TARGET_GH_TOKEN: ${{ (github.event.inputs.target_repo == '' || github.event.inputs.target_repo == github.repository) && secrets.GITHUB_TOKEN || secrets.GH_PAT_WORKFLOWS }}
+      TARGET_GH_TOKEN: ${{ (github.event.inputs.target_repo == '' || github.event.inputs.target_repo == github.repository) && secrets.GITHUB_TOKEN || secrets.GH_PAT_DON_PETRY || secrets.GH_PAT_WORKFLOWS }}
       DISCUSSION_NUMBER: ${{ github.event.inputs.discussion }}
       DRY_RUN: ${{ inputs.dry_run == true && '1' || '0' }}
       FORCE_REPLAN: ${{ inputs.force_replan == true && '1' || '0' }}
diff --git a/.github/workflows/lint.yml b/.github/workflows/lint.yml
index 9c1a5011a..fc8814a92 100644
--- a/.github/workflows/lint.yml
+++ b/.github/workflows/lint.yml
@@ -334,9 +334,28 @@ jobs:
 
       - name: Verify live persona addressing teams
         env:
-          GH_TOKEN: ${{ secrets.GH_PAT_WORKFLOWS }}
+          GH_TOKEN: ${{ secrets.GH_PAT_DON_PETRY || secrets.GH_PAT_WORKFLOWS }}
         run: bash scripts/verify-persona-teams.sh personas
 
+  verify-persona-identity:
+    # Drift check (.github-private#1316) that ties each persona's declared
+    # runtime.identity to the workflows that consume it: the credential follows the
+    # GH_PAT_<ACCOUNT> naming convention, the secret is actually referenced by a
+    # workflow, and every persona shipping a write workflow resolves BOT_USER from
+    # the manifest (not the shared vars.BOT_USER that let dev-lead regress to
+    # committing as donpetry-bot). No GitHub API calls / no gh — pure
+    # manifest↔workflow consistency (PyYAML installed via pip below).
+    runs-on: ubuntu-latest
+    timeout-minutes: 5
+    steps:
+      - uses: actions/checkout@df4cb1c069e1874edd31b4311f1884172cec0e10  # v6.0.3
+
+      - name: Install PyYAML
+        run: pip3 install --quiet pyyaml
+
+      - name: Verify persona identity ↔ workflow wiring
+        run: bash scripts/verify-persona-identity.sh personas
+
   gh-aw-compile:
     runs-on: ubuntu-latest
     timeout-minutes: 5
diff --git a/.github/workflows/persona-runner-reusable.yml b/.github/workflows/persona-runner-reusable.yml
index f948dfdeb..adb41e6c5 100644
--- a/.github/workflows/persona-runner-reusable.yml
+++ b/.github/workflows/persona-runner-reusable.yml
@@ -53,7 +53,12 @@ on:
     secrets:
       CLAUDE_CODE_OAUTH_TOKEN:
         required: true
+      GH_PAT_DON_PETRY:
+        # don-petry's PAT — the account the persona comment-poster authenticates as.
+        # Canonical name (GH_PAT_<ACCOUNT> schema); supersedes GH_PAT_WORKFLOWS below.
+        required: false
       GH_PAT_WORKFLOWS:
+        # DEPRECATED alias of GH_PAT_DON_PETRY — kept during the rename migration.
         required: true
 
 permissions: {}
@@ -179,7 +184,7 @@ jobs:
         if: steps.resolve.outputs.skip == 'false' && steps.advise.outputs.skip_post == 'false'
         env:
           # The PAT lives ONLY here. The agent step above never had it.
-          GH_TOKEN: ${{ secrets.GH_PAT_WORKFLOWS }}
+          GH_TOKEN: ${{ secrets.GH_PAT_DON_PETRY || secrets.GH_PAT_WORKFLOWS }}
           PERSONA: ${{ inputs.persona }}
           SOURCE_REPO: ${{ inputs.source_repo }}
           ITEM_NUMBER: ${{ inputs.item_number }}
diff --git a/.github/workflows/pr-review-canary.yml b/.github/workflows/pr-review-canary.yml
index 04f34fd62..5d2539cd0 100644
--- a/.github/workflows/pr-review-canary.yml
+++ b/.github/workflows/pr-review-canary.yml
@@ -59,7 +59,7 @@ jobs:
       - name: Run pr-review-trigger canary (dry-run)
         env:
           # PAT required — a workflow_dispatch fired with GITHUB_TOKEN never starts a run.
-          GH_TOKEN: ${{ secrets.GH_PAT_WORKFLOWS }}
+          GH_TOKEN: ${{ secrets.GH_PAT_DON_PETRY || secrets.GH_PAT_WORKFLOWS }}
           REPO: ${{ github.repository }}
           CANARY_OUT: pr_review_canary.md
         run: bash scripts/pr_review_canary.sh
diff --git a/.github/workflows/pr-review-mention.yml b/.github/workflows/pr-review-mention.yml
index dfab4345a..d07129aa3 100644
--- a/.github/workflows/pr-review-mention.yml
+++ b/.github/workflows/pr-review-mention.yml
@@ -39,5 +39,5 @@ jobs:
       pull-requests: write
     uses: petry-projects/.github/.github/workflows/pr-review-mention-reusable.yml@pr-review-mention/v2-next
     secrets:
-      GH_PAT_WORKFLOWS: ${{ secrets.GH_PAT_WORKFLOWS }}
+      GH_PAT_WORKFLOWS: ${{ secrets.GH_PAT_DON_PETRY || secrets.GH_PAT_WORKFLOWS }}
       DON_PETRY_BOT_GH_PAT: ${{ secrets.DON_PETRY_BOT_GH_PAT }}
diff --git a/.github/workflows/pr-review-sweep.yml b/.github/workflows/pr-review-sweep.yml
index 81b74b369..10d5d45bf 100644
--- a/.github/workflows/pr-review-sweep.yml
+++ b/.github/workflows/pr-review-sweep.yml
@@ -84,7 +84,7 @@ jobs:
         # the sweep "succeeds" while re-triggering nothing (the original #573 gap
         # regression). Fail loudly instead of silently no-op'ing.
         env:
-          GH_TOKEN: ${{ secrets.GH_PAT_WORKFLOWS }}
+          GH_TOKEN: ${{ secrets.GH_PAT_DON_PETRY || secrets.GH_PAT_WORKFLOWS }}
         run: |
           if [ -z "${GH_TOKEN}" ]; then
             echo "::error::GH_PAT_WORKFLOWS is required — GITHUB_TOKEN cannot start a dispatched run and the bot comment PAT lacks actions:write."
@@ -96,7 +96,7 @@ jobs:
           # PAT with workflow/actions:write scope required: GITHUB_TOKEN cannot
           # start dispatched workflow runs, and the bot comment PAT lacks the
           # actions:write scope (HTTP 403 on the dispatch).
-          GH_TOKEN: ${{ secrets.GH_PAT_WORKFLOWS }}
+          GH_TOKEN: ${{ secrets.GH_PAT_DON_PETRY || secrets.GH_PAT_WORKFLOWS }}
           DRY_RUN: ${{ inputs.dry_run || 'false' }}
           MAX_DISPATCH: ${{ inputs.max_dispatch || '20' }}
         run: bash scripts/sweep-stuck-reviews.sh
diff --git a/.github/workflows/pr-review.yml b/.github/workflows/pr-review.yml
index a1b6f0158..be9aee571 100644
--- a/.github/workflows/pr-review.yml
+++ b/.github/workflows/pr-review.yml
@@ -46,7 +46,13 @@ on:
       # over DON_PETRY_BOT_GH_PAT when set; see docs/pr-review-agent/machine-user-setup.md.
       DON_PETRY_BOT_GH_PAT_CLASSIC:
         required: false
+      GH_PAT_DON_PETRY_COPILOT:
+        # don-petry's Copilot-subscription PAT (Copilot review engine). Canonical
+        # name under the GH_PAT_<ACCOUNT> schema; supersedes GH_PAT (below, kept as
+        # a transition fallback). The bot's own PATs keep their grandfathered names.
+        required: false
       GH_PAT:
+        # DEPRECATED alias of GH_PAT_DON_PETRY_COPILOT — kept during migration.
         required: false
 
 # ── Reusable-only (#497 self-host pinning) ───────────────────────────────
@@ -138,7 +144,7 @@ jobs:
       REVIEW_ENGINE: ${{ vars.REVIEW_ENGINE || 'claude' }}
       CLAUDE_CODE_OAUTH_TOKEN: ${{ secrets.CLAUDE_CODE_OAUTH_TOKEN }}
       # Copilot engine reuses GH_PAT (a separate user PAT with Copilot subscription).
-      COPILOT_GITHUB_TOKEN: ${{ secrets.GH_PAT }}
+      COPILOT_GITHUB_TOKEN: ${{ secrets.GH_PAT_DON_PETRY_COPILOT || secrets.GH_PAT }}
       # Gemini engine auth — GOOGLE_API_KEY is the secret name; newer Gemini CLI
       # versions expect GEMINI_API_KEY so we expose the same value under both names.
       GOOGLE_API_KEY: ${{ secrets.GOOGLE_API_KEY }}
@@ -211,6 +217,20 @@ jobs:
           repository: petry-projects/.github-private
           ref: ${{ inputs.agent_ref || 'main' }}
 
+      - name: Resolve persona identity
+        # pr-review's acting account is DECLARED in personas/pr-review/persona.yml
+        # (runtime.identity.account = donpetry-bot, the review-only machine user).
+        # Resolving BOT_USER from the manifest — instead of the shared vars.BOT_USER
+        # — is what keeps this identity from colliding with dev-lead's (#1316). The
+        # job-env BOT_USER remains the fail-safe default if the manifest is unreadable.
+        run: |
+          if account="$(bash scripts/lib/resolve-persona-identity.sh pr-review personas account)"; then
+            echo "BOT_USER=$account" >> "$GITHUB_ENV"
+            echo "resolved pr-review identity: BOT_USER=$account"
+          else
+            echo "::warning::could not resolve pr-review identity from manifest; using fallback BOT_USER=$BOT_USER"
+          fi
+
       - name: Cache claude-code CLI
         uses: actions/cache@27d5ce7f107fe9357f9df03efb73ab90386fccae # v5.0.5
         with:
diff --git a/.github/workflows/readme-refresh.yml b/.github/workflows/readme-refresh.yml
index 17775c837..1595d3ba4 100644
--- a/.github/workflows/readme-refresh.yml
+++ b/.github/workflows/readme-refresh.yml
@@ -65,6 +65,6 @@ jobs:
 
       - name: Run README refresh
         env:
-          GH_TOKEN: ${{ secrets.GH_PAT }}
+          GH_TOKEN: ${{ secrets.GH_PAT_DON_PETRY_COPILOT || secrets.GH_PAT }}
           BOT_USER: don-petry   # commit/push as don-petry (GH_PAT owner), not donpetry-bot
         run: bash scripts/aw-readme-refresh.sh
diff --git a/.github/workflows/reviewer-report.yml b/.github/workflows/reviewer-report.yml
index 4df8b48ac..687a7a869 100644
--- a/.github/workflows/reviewer-report.yml
+++ b/.github/workflows/reviewer-report.yml
@@ -59,11 +59,11 @@ jobs:
         with:
           repository: petry-projects/.github-private
           ref: main
-          token: ${{ secrets.GH_PAT_WORKFLOWS }}
+          token: ${{ secrets.GH_PAT_DON_PETRY || secrets.GH_PAT_WORKFLOWS }}
 
       - name: Generate org-wide reviewer scorecard
         env:
-          GH_TOKEN: ${{ secrets.GH_PAT_WORKFLOWS }}
+          GH_TOKEN: ${{ secrets.GH_PAT_DON_PETRY || secrets.GH_PAT_WORKFLOWS }}
           REVIEWER_REPORT_OUT: reviewer_report.md
           REVIEWER_SNAPSHOT_OUT: reviewer-report-state.json
         run: |
diff --git a/.github/workflows/token-report.yml b/.github/workflows/token-report.yml
index b2c6b4ee4..dc15c3ed1 100644
--- a/.github/workflows/token-report.yml
+++ b/.github/workflows/token-report.yml
@@ -56,11 +56,11 @@ jobs:
         with:
           repository: petry-projects/.github-private
           ref: main
-          token: ${{ secrets.GH_PAT_WORKFLOWS }}
+          token: ${{ secrets.GH_PAT_DON_PETRY || secrets.GH_PAT_WORKFLOWS }}
 
       - name: Generate org-wide token report
         env:
-          GH_TOKEN: ${{ secrets.GH_PAT_WORKFLOWS }}
+          GH_TOKEN: ${{ secrets.GH_PAT_DON_PETRY || secrets.GH_PAT_WORKFLOWS }}
           TOKEN_REPORT_OUT: token_report.md
         run: |
           bash scripts/token_report.sh >> "$GITHUB_STEP_SUMMARY"
diff --git a/AGENTS.md b/AGENTS.md
index 43987a67c..b6bb85e56 100644
--- a/AGENTS.md
+++ b/AGENTS.md
@@ -302,6 +302,27 @@ run to `remediate=live` for the configured pilot.
 - Scripts must be POSIX-compatible shell (`#!/usr/bin/env bash` with `set -euo pipefail`).
 - No hardcoded tokens or secrets — use `$GITHUB_TOKEN` from the environment.
 
+### Agent identity & credential secrets
+
+Every persona/agent declares the GitHub account it **acts as** — the login it
+commits, pushes, reviews, or comments as — in its manifest under
+`runtime.identity` (`personas/<id>/persona.yml`): an `account` and the `credential`
+secret holding that account's PAT. The runtime resolves `BOT_USER` from
+`account` via `scripts/lib/resolve-persona-identity.sh`; **never** derive a
+write persona's identity from a shared `vars.BOT_USER` — that coupling let
+dev-lead regress to committing as the review-only machine user `donpetry-bot`
+([#1316](https://github.com/petry-projects/.github-private/issues/1316)).
+
+- **dev-lead** acts as `don-petry` (the owner); **pr-review** acts as the
+  review-only machine user `donpetry-bot`. They are deliberately different accounts.
+- PAT secrets follow **`GH_PAT_<ACCOUNT>[_<QUALIFIER>]`** (login upper-snake-cased),
+  e.g. `GH_PAT_DON_PETRY`, `GH_PAT_DON_PETRY_COPILOT`. Pre-existing account-named
+  secrets (`DON_PETRY_BOT_GH_PAT`) are grandfathered.
+- `lint.yml`'s `verify-persona-identity` job enforces the manifest↔workflow
+  wiring; `validate-personas.py` makes `runtime.identity` mandatory. See
+  `standards/persona-standards.md` §5.1 and
+  `docs/pr-review-agent/machine-user-setup.md`.
+
 ### Initiative Planner — blocking open-questions gate
 
 `scripts/initiative-planner/apply-plan.sh` will **not** materialize an epic + sub-issue
diff --git a/docs/dev-lead/spec.md b/docs/dev-lead/spec.md
index 8bb9935a3..6f4346882 100644
--- a/docs/dev-lead/spec.md
+++ b/docs/dev-lead/spec.md
@@ -290,9 +290,13 @@ env:
   CLAUDE_CODE_OAUTH_TOKEN: ${{ secrets.CLAUDE_CODE_OAUTH_TOKEN }}
   GOOGLE_API_KEY: ${{ secrets.GOOGLE_API_KEY }}
   GEMINI_API_KEY: ${{ secrets.GOOGLE_API_KEY }}
-  COPILOT_GITHUB_TOKEN: ${{ secrets.GH_PAT }}
-  GH_TOKEN: ${{ secrets.GH_PAT_WORKFLOWS || github.token }}
-  BOT_USER: ${{ vars.BOT_USER || 'donpetry-bot' }}
+  COPILOT_GITHUB_TOKEN: ${{ secrets.GH_PAT_DON_PETRY_COPILOT || secrets.GH_PAT }}
+  GH_TOKEN: ${{ secrets.GH_PAT_DON_PETRY || secrets.GH_PAT_WORKFLOWS || github.token }}
+  # Hardcoded fail-safe default only; the real value is resolved from
+  # personas/dev-lead/persona.yml (runtime.identity.account = don-petry) by the
+  # "Resolve persona identity" step. vars.BOT_USER is NOT consulted — that
+  # shared variable caused the donpetry-bot regression (#1316). See persona-standards.md §5.1.
+  BOT_USER: 'don-petry'
   TRUSTED_BOTS: ${{ vars.TRUSTED_BOTS || 'copilot-pull-request-reviewer[bot],gemini-code-assist[bot],coderabbitai[bot],sonarqubecloud[bot]' }}
   TRIGGER_PHRASES: ${{ vars.TRIGGER_PHRASES || '@claude,@dev-lead' }}
 ```
@@ -579,7 +583,7 @@ No external script. Runs entirely as a `run:` shell block within the job.
 | `CLAUDE_CODE_VERSION` | `latest` | Claude CLI version to install and cache |
 | `TRUSTED_BOTS` | See §5.2 | Comma-separated list of trusted bot logins |
 | `TRIGGER_PHRASES` | `@claude,@dev-lead` | Phrases that trigger human intent in comments |
-| `BOT_USER` | `donpetry-bot` | Machine user login (excluded from self-triggering) |
+| `BOT_USER` | `don-petry` (resolved from `personas/dev-lead/persona.yml` `runtime.identity.account`) | The account dev-lead acts as — commits/pushes and self-exclusion. Manifest-driven since #1316; the env value is only a fail-safe default. |
 | `MAX_CI_CYCLES` | `3` | Max fix-CI loops before giving up |
 | `MAX_REVIEW_CYCLES` | `3` | Max review-fix loops before giving up |
 | `DEV_LEAD_LIVE_MODE` | `false` | If `true`, run live on all events; if `false`, dry-run for scheduled (N/A for event-driven) |
diff --git a/docs/pr-review-agent/machine-user-setup.md b/docs/pr-review-agent/machine-user-setup.md
index 78e507e24..f3c4bd373 100644
--- a/docs/pr-review-agent/machine-user-setup.md
+++ b/docs/pr-review-agent/machine-user-setup.md
@@ -118,11 +118,26 @@ gh secret list --repo petry-projects/.github-private | grep DON_PETRY_BOT_GH_PAT
 
 ### Secret naming reference
 
-| Secret | Purpose |
-|--------|---------|
-| `DON_PETRY_BOT_GH_PAT` | Machine user / bot reviewer PAT. Used by `pr-review.yml`, `repair-pr-approvals.yml`, `daily-pr-review-health.yml`, and `claude.yml` for GitHub API authentication as the bot. |
-| `GH_PAT` | User PAT with a Copilot subscription. Exposed as `COPILOT_GITHUB_TOKEN` in `pr-review.yml` for Copilot CLI access. Do **not** store the bot reviewer PAT here. |
-| `GOOGLE_API_KEY` | API key for Google AI (Gemini). Used by the `gemini` engine for triage, deep review, and audit tiers. |
+PAT secrets follow the schema **`GH_PAT_<ACCOUNT>[_<QUALIFIER>]`**, where `<ACCOUNT>`
+is the GitHub login upper-snake-cased (`don-petry` → `GH_PAT_DON_PETRY`). This makes
+the *owning account* obvious from the secret name, resolving the ambiguity that let dev-lead
+regress to committing as the bot ([#1316](https://github.com/petry-projects/.github-private/issues/1316)).
+Each persona declares which account/secret it acts as in `runtime.identity`
+(`personas/<id>/persona.yml`); see `standards/persona-standards.md` §5.1.
+
+| Secret | Account | Purpose |
+|--------|---------|---------|
+| `GH_PAT_DON_PETRY` | don-petry | Write/workflows PAT — the account dev-lead and the persona runner commit/push/comment as. **Canonical name** for the PAT formerly stored as `GH_PAT_WORKFLOWS`. |
+| `GH_PAT_DON_PETRY_COPILOT` | don-petry | User PAT with a Copilot subscription. Exposed as `COPILOT_GITHUB_TOKEN` for Copilot CLI access. **Canonical name** for the PAT formerly stored as `GH_PAT`. |
+| `DON_PETRY_BOT_GH_PAT` | donpetry-bot | Machine user / bot reviewer PAT. Used by `pr-review.yml`, `repair-pr-approvals.yml`, `daily-pr-review-health.yml` for GitHub API auth as the bot. Grandfathered name (pre-dates the schema). |
+| `DON_PETRY_BOT_GH_PAT_CLASSIC` | donpetry-bot | Classic PAT for the bot; required for `gh pr review --approve`. Preferred over the fine-grained secret when set. Grandfathered. |
+| `GOOGLE_API_KEY` | — | API key for Google AI (Gemini). Used by the `gemini` engine for triage, deep review, and audit tiers. |
+
+> **Migration:** `GH_PAT_WORKFLOWS` → `GH_PAT_DON_PETRY` and `GH_PAT` →
+> `GH_PAT_DON_PETRY_COPILOT`. Workflows read `${{ secrets.GH_PAT_DON_PETRY || secrets.GH_PAT_WORKFLOWS }}`
+> during the transition, so **create the new secrets as copies of the old ones**;
+> once confirmed present, a follow-up drops the `|| secrets.GH_PAT_WORKFLOWS`
+> fallbacks and deletes the old secrets.
 
 ## Step 5: Verify the setup
 
diff --git a/personas/business-analyst/persona.yml b/personas/business-analyst/persona.yml
index c7bee950d..90cbb5896 100644
--- a/personas/business-analyst/persona.yml
+++ b/personas/business-analyst/persona.yml
@@ -67,6 +67,15 @@ triggers:
 trust:
   author_association_floor: [OWNER, MEMBER, COLLABORATOR]
 
+# --- Runtime identity (§5.1) ------------------------------------------------
+# Advisory-only (producers under prompts/incubation, consumed by path and the
+# persona runner). On the comment path it posts as `don-petry` (the runner
+# authenticates with that account's PAT). No dedicated reusable yet.
+runtime:
+  identity:
+    account: don-petry
+    credential: GH_PAT_DON_PETRY
+
 evals:
   required_before: stable
   path: evals/business-analyst/
diff --git a/personas/dev-lead/persona.yml b/personas/dev-lead/persona.yml
index e77dcffb1..ab95a1096 100644
--- a/personas/dev-lead/persona.yml
+++ b/personas/dev-lead/persona.yml
@@ -83,6 +83,19 @@ triggers:
 trust:
   author_association_floor: [OWNER, MEMBER, COLLABORATOR]
 
+# --- Runtime identity (§5.1) ------------------------------------------------
+# The account dev-lead ACTS AS: it implements issues into PRs and commits/pushes
+# fixes, so it must author commits as the human owner `don-petry`, NOT the
+# review-only machine user `donpetry-bot`. dev-lead-reusable.yml resolves
+# BOT_USER from identity.account (see scripts/lib/resolve-persona-identity.sh);
+# git-identity.sh then attributes commits to it. Distinct from address.handle
+# (petry-projects/dev-lead, the mention team). This decoupling is the fix for
+# the donpetry-bot regression (.github-private#1316).
+runtime:
+  identity:
+    account: don-petry
+    credential: GH_PAT_DON_PETRY
+
 evals:
   required_before: stable
   path: evals/dev-lead/
diff --git a/personas/devops-lead/persona.yml b/personas/devops-lead/persona.yml
index 753b1e14b..09c15f879 100644
--- a/personas/devops-lead/persona.yml
+++ b/personas/devops-lead/persona.yml
@@ -72,9 +72,14 @@ triggers:
 trust:
   author_association_floor: [OWNER, MEMBER, COLLABORATOR]
 
-# No `runtime:` block yet — devops-lead ships no dedicated reusable/stub in this
-# pilot. It is consumed by path and is human-invocable. Add `runtime` + a
-# canary-rings.json entry when a dedicated advisory workflow is wired.
+# --- Runtime identity (§5.1) ------------------------------------------------
+# Advisory-only: devops-lead ships no dedicated reusable/stub in this pilot.
+# It is consumed by path and is human-invocable. On the comment path it posts
+# as `don-petry` (the persona runner authenticates with that account's PAT).
+runtime:
+  identity:
+    account: don-petry
+    credential: GH_PAT_DON_PETRY
 
 evals:
   required_before: stable
diff --git a/personas/pr-review/persona.yml b/personas/pr-review/persona.yml
index 553bb6c97..8146cf9fc 100644
--- a/personas/pr-review/persona.yml
+++ b/personas/pr-review/persona.yml
@@ -64,6 +64,18 @@ triggers:
 trust:
   author_association_floor: [OWNER, MEMBER, COLLABORATOR]
 
+# --- Runtime identity (§5.1) ------------------------------------------------
+# pr-review ACTS AS the dedicated machine user `donpetry-bot`: a review-only
+# account distinct from `don-petry` so its approvals satisfy CODEOWNERS and it
+# never authors code. It posts NO commits (review-only) — the identity is used
+# for its API auth and approvals. Credential is the bot's classic PAT (required
+# for addPullRequestReview); the base secret name is grandfathered (pre-dates
+# the GH_PAT_<ACCOUNT> convention). See docs/pr-review-agent/machine-user-setup.md.
+runtime:
+  identity:
+    account: donpetry-bot
+    credential: DON_PETRY_BOT_GH_PAT
+
 # pr-review wraps two skills (deep-review + triage); their cases are combined
 # under evals/pr-review/ to satisfy the schema's single-path constraint.
 evals:
diff --git a/personas/qa-lead/persona.yml b/personas/qa-lead/persona.yml
index 3051cfdff..5d3af0600 100644
--- a/personas/qa-lead/persona.yml
+++ b/personas/qa-lead/persona.yml
@@ -122,10 +122,16 @@ triggers:
 trust:
   author_association_floor: [OWNER, MEMBER, COLLABORATOR]
 
-# No `runtime:` block yet — qa-lead ships no dedicated reusable/stub in this
-# pilot. It is consumed by path (e.g. by the initiative-planner during
-# planning) and is human-invocable. Add `runtime` + a canary-rings.json entry
-# when a dedicated advisory workflow is wired.
+# qa-lead ships no dedicated reusable/stub in this pilot — it is consumed by
+# path (e.g. by the initiative-planner) and is human-invocable via the persona
+# runner. It declares only `runtime.identity`: on the advisory/comment path it
+# posts as `don-petry` (the persona runner authenticates with that account's
+# PAT). Add the rest of `runtime` + a canary-rings.json entry when a dedicated
+# advisory workflow is wired.
+runtime:
+  identity:
+    account: don-petry
+    credential: GH_PAT_DON_PETRY
 
 evals:
   required_before: stable
diff --git a/personas/scrum-master/persona.yml b/personas/scrum-master/persona.yml
index b836a5374..fd42899cc 100644
--- a/personas/scrum-master/persona.yml
+++ b/personas/scrum-master/persona.yml
@@ -69,6 +69,15 @@ triggers:
 trust:
   author_association_floor: [OWNER, MEMBER, COLLABORATOR]
 
+# --- Runtime identity (§5.1) ------------------------------------------------
+# Advisory-only (consumed by the initiative-planner and the persona runner). On
+# the comment path it posts as `don-petry` (the runner authenticates with that
+# account's PAT). No dedicated reusable yet.
+runtime:
+  identity:
+    account: don-petry
+    credential: GH_PAT_DON_PETRY
+
 evals:
   required_before: stable
   path: evals/scrum-master/
diff --git a/personas/security-lead/persona.yml b/personas/security-lead/persona.yml
index 981df0822..b6666798a 100644
--- a/personas/security-lead/persona.yml
+++ b/personas/security-lead/persona.yml
@@ -86,9 +86,14 @@ triggers:
 trust:
   author_association_floor: [OWNER, MEMBER, COLLABORATOR]
 
-# No `runtime:` block yet — security-lead ships no dedicated reusable/stub in
-# this pilot. It is consumed by path and is human-invocable. Add `runtime` + a
-# canary-rings.json entry when a dedicated advisory workflow is wired.
+# --- Runtime identity (§5.1) ------------------------------------------------
+# Advisory-only: security-lead ships no dedicated reusable/stub in this pilot.
+# It is consumed by path and is human-invocable. On the comment path it posts
+# as `don-petry` (the persona runner authenticates with that account's PAT).
+runtime:
+  identity:
+    account: don-petry
+    credential: GH_PAT_DON_PETRY
 
 evals:
   required_before: stable
diff --git a/personas/sre-lead/persona.yml b/personas/sre-lead/persona.yml
index e160266ef..439cda78d 100644
--- a/personas/sre-lead/persona.yml
+++ b/personas/sre-lead/persona.yml
@@ -84,9 +84,14 @@ triggers:
 trust:
   author_association_floor: [OWNER, MEMBER, COLLABORATOR]
 
-# No `runtime:` block yet — sre-lead ships no dedicated reusable/stub in this
-# pilot. It is consumed by path and is human-invocable. Add `runtime` + a
-# canary-rings.json entry when a dedicated advisory workflow is wired.
+# --- Runtime identity (§5.1) ------------------------------------------------
+# Advisory-only: sre-lead ships no dedicated reusable/stub in this pilot.
+# It is consumed by path and is human-invocable. On the comment path it posts
+# as `don-petry` (the persona runner authenticates with that account's PAT).
+runtime:
+  identity:
+    account: don-petry
+    credential: GH_PAT_DON_PETRY
 
 evals:
   required_before: stable
diff --git a/personas/validate-personas.py b/personas/validate-personas.py
index af21186a4..2f7b9ad5a 100644
--- a/personas/validate-personas.py
+++ b/personas/validate-personas.py
@@ -188,6 +188,39 @@ def check_evals(manifest: dict, manifest_path: Path, repo_root: Path) -> None:
                      f"placeholder evals)")
 
 
+# Credentials that pre-date the GH_PAT_<ACCOUNT> naming convention and are kept
+# as-is (they are already account-named, just in the legacy suffix order). New
+# personas MUST use the GH_PAT_<ACCOUNT>[_<QUALIFIER>] form.
+GRANDFATHERED_CREDENTIALS = frozenset({
+    "DON_PETRY_BOT_GH_PAT",
+    "DON_PETRY_BOT_GH_PAT_CLASSIC",
+})
+
+
+def check_identity(manifest: dict, manifest_path: Path) -> None:
+    """Every persona MUST declare the account it acts as. The schema validates the
+    SHAPE of runtime.identity when present; this makes it MANDATORY and enforces
+    the credential naming convention — so no persona can silently fall back to a
+    shared machine user (the dev-lead->donpetry-bot regression, .github-private#1316)."""
+    runtime = manifest.get("runtime")
+    if not isinstance(runtime, dict) or "identity" not in runtime:
+        fail(f"{manifest_path}: missing runtime.identity — every persona must declare the "
+             f"account it acts as (account + credential). See persona-standards.md §5.1.")
+    identity = runtime["identity"]
+    if not isinstance(identity, dict) or "credential" not in identity or "account" not in identity:
+        fail(f"{manifest_path}: runtime.identity must be a mapping containing 'account' and "
+             f"'credential'. See persona-standards.md §5.1.")
+    credential = identity["credential"]
+    account = identity["account"]
+    if credential not in GRANDFATHERED_CREDENTIALS:
+        upper_account = account.upper().replace("-", "_")
+        prefix = f"GH_PAT_{upper_account}"
+        if credential != prefix and not credential.startswith(prefix + "_"):
+            fail(f"{manifest_path}: runtime.identity.credential '{credential}' must follow the "
+                 f"GH_PAT_<ACCOUNT>[_<QUALIFIER>] convention for account '{account}' "
+                 f"(expected '{prefix}[_<QUALIFIER>]'). See persona-standards.md §5.1.")
+
+
 def check_invariants(manifest: dict, manifest_path: Path, repo_root: Path,
                      registry_arg: str | None) -> None:
     persona_dir = manifest_path.parent
@@ -226,6 +259,8 @@ def check_invariants(manifest: dict, manifest_path: Path, repo_root: Path,
                 fail(f"{manifest_path}: vendor_pin '{fw['vendor_pin']}' not found in {vendor_md} "
                      f"(manifest pin and vendored version disagree)")
 
+    check_identity(manifest, manifest_path)
+
     check_evals(manifest, manifest_path, repo_root)
 
     if manifest["status"] != "draft":
diff --git a/scripts/lib/resolve-persona-identity.sh b/scripts/lib/resolve-persona-identity.sh
new file mode 100755
index 000000000..7dbabae8f
--- /dev/null
+++ b/scripts/lib/resolve-persona-identity.sh
@@ -0,0 +1,94 @@
+#!/usr/bin/env bash
+# resolve-persona-identity.sh — print a persona's declared runtime identity.
+#
+# Usage: resolve-persona-identity.sh <persona-id> <personas-root> [field]
+#   field: account (default) | credential
+#
+# Reads <personas-root>/<persona-id>/persona.yml and prints runtime.identity.<field>.
+# This is the SINGLE runtime source of a persona's acting GitHub identity: the
+# dev-lead and pr-review workflows resolve BOT_USER from `account` here instead of
+# the shared `vars.BOT_USER` that let dev-lead regress to committing as the
+# review-only machine user donpetry-bot (.github-private#1316).
+#
+# Exits non-zero (message on stderr) if the manifest, the identity block, or the
+# requested field is missing — it never emits a silent fallback identity.  Callers
+# decide whether to treat the non-zero exit as fatal or fall back to a safe default.
+set -euo pipefail
+
+id="${1:?usage: resolve-persona-identity.sh <persona-id> <personas-root> [field]}"
+root="${2:?usage: resolve-persona-identity.sh <persona-id> <personas-root> [field]}"
+field="${3:-account}"
+
+case "$field" in
+  account | credential) ;;
+  *)
+    echo "resolve-persona-identity: field must be 'account' or 'credential', got '$field'" >&2
+    exit 2
+    ;;
+esac
+
+manifest="${root%/}/${id}/persona.yml"
+if [ ! -f "$manifest" ]; then
+  echo "resolve-persona-identity: no manifest at $manifest" >&2
+  exit 1
+fi
+
+# Prefer PyYAML (same parser validate-personas.py uses); fall back to a simple
+# standard-library-only parser if PyYAML is not installed to avoid pip install
+# overhead and PEP 668 ("externally-managed-environment") failures.
+python3 - "$manifest" "$field" <<'PY'
+import sys
+
+manifest, field = sys.argv[1], sys.argv[2]
+val = None
+
+try:
+    import yaml
+    with open(manifest, encoding="utf-8") as fh:
+        doc = yaml.safe_load(fh) or {}
+    identity = ((doc.get("runtime") or {}).get("identity")) or {}
+    val = identity.get(field)
+except ImportError:
+    # Fallback to a simple line-by-line parser if PyYAML is not installed
+    try:
+        with open(manifest, encoding="utf-8") as fh:
+            state = "start"
+            runtime_indent = -1
+            identity_indent = -1
+            for line in fh:
+                stripped = line.strip()
+                if not stripped or stripped.startswith("#"):
+                    continue
+                indent = len(line) - len(line.lstrip())
+                if state == "start" and stripped == "runtime:":
+                    state = "runtime"
+                    runtime_indent = indent
+                elif state == "runtime":
+                    if indent <= runtime_indent:
+                        state = "start"
+                    elif stripped == "identity:":
+                        state = "identity"
+                        identity_indent = indent
+                elif state == "identity":
+                    if indent <= identity_indent:
+                        if indent <= runtime_indent:
+                            state = "start"
+                        else:
+                            state = "runtime"
+                    else:
+                        if ":" in stripped:
+                            k, v = stripped.split(":", 1)
+                            if k.strip() == field:
+                                val = v.strip().strip("'\"")
+                                break
+    except Exception as e:
+        sys.stderr.write(f"resolve-persona-identity: fallback parser failed: {e}\n")
+        sys.exit(5)
+
+if not val:
+    sys.stderr.write(
+        f"resolve-persona-identity: runtime.identity.{field} missing in {manifest}\n"
+    )
+    sys.exit(4)
+print(val)
+PY
diff --git a/scripts/verify-persona-identity.sh b/scripts/verify-persona-identity.sh
new file mode 100755
index 000000000..a3057df59
--- /dev/null
+++ b/scripts/verify-persona-identity.sh
@@ -0,0 +1,99 @@
+#!/usr/bin/env bash
+# verify-persona-identity.sh — LIVE-ish drift check for persona runtime identity.
+#
+# The schema (persona.schema.json) validates the SHAPE of runtime.identity and
+# validate-personas.py makes it MANDATORY. This script closes the loop between the
+# manifest and the workflows that consume it, so the declared identity and the
+# actual CI wiring can never silently diverge (the failure mode behind
+# .github-private#1316). For each persona it asserts:
+#
+#   1. credential naming — a non-grandfathered credential MUST be
+#      GH_PAT_<ACCOUNT_UPPER_SNAKE>[_<QUALIFIER>] for the declared account.
+#   2. credential is real — the secret is referenced by at least one workflow
+#      (guards against declaring a phantom/typo'd secret).
+#   3. runtime consumption — a persona that ships a write workflow MUST resolve
+#      its identity from the manifest (call resolve-persona-identity.sh), not from
+#      the shared vars.BOT_USER.
+#
+# Usage: verify-persona-identity.sh [personas-root]
+# Exit 0 = all good; exit 1 = a drift/violation was found.
+set -euo pipefail
+
+ROOT="${1:-personas}"
+SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+RESOLVE="$SCRIPT_DIR/lib/resolve-persona-identity.sh"
+WORKFLOWS_DIR=".github/workflows"
+
+# Credentials that pre-date the GH_PAT_<ACCOUNT> convention (kept as-is). Mirror
+# GRANDFATHERED_CREDENTIALS in personas/validate-personas.py.
+GRANDFATHERED=(
+  DON_PETRY_BOT_GH_PAT
+  DON_PETRY_BOT_GH_PAT_CLASSIC
+)
+
+# Personas that ship a write workflow which MUST consume the manifest identity.
+# id:workflow-path pairs. Advisory personas post via the shared persona runner
+# and are covered by checks 1-2 only.
+declare -a WRITE_RUNTIMES=(
+  "dev-lead:.github/workflows/dev-lead-reusable.yml"
+  "pr-review:.github/workflows/pr-review.yml"
+)
+
+fail() { echo "::error::verify-persona-identity: $*" >&2; exit 1; }
+
+is_grandfathered() {
+  local c="$1" g
+  for g in "${GRANDFATHERED[@]}"; do [ "$c" = "$g" ] && return 0; done
+  return 1
+}
+
+# Upper-snake-case conversion for account names is done inline via pure Bash parameter expansion.
+
+runtime_for() {
+  local id="$1" pair
+  for pair in "${WRITE_RUNTIMES[@]}"; do
+    [ "${pair%%:*}" = "$id" ] && { printf '%s' "${pair#*:}"; return 0; }
+  done
+  return 1
+}
+
+count=0
+for manifest in "$ROOT"/*/persona.yml; do
+  [ -f "$manifest" ] || continue
+  manifest_dir="${manifest%/persona.yml}"
+  id="${manifest_dir##*/}"
+  account="$(bash "$RESOLVE" "$id" "$ROOT" account)"
+  credential="$(bash "$RESOLVE" "$id" "$ROOT" credential)"
+
+  # 1. credential naming convention
+  if ! is_grandfathered "$credential"; then
+    upper_account="${account^^}"
+    expect="GH_PAT_${upper_account//-/_}"
+    case "$credential" in
+      "$expect" | "${expect}_"*) : ;;
+      *) fail "$id: credential '$credential' does not match '${expect}[_<QUALIFIER>]' for account '$account'" ;;
+    esac
+  fi
+
+  # 2. credential is referenced by at least one workflow
+  if ! grep -rqlE "secrets\.${credential}([^A-Z0-9_]|$)" "$WORKFLOWS_DIR" 2>/dev/null; then
+    fail "$id: credential secret '$credential' is not referenced by any workflow in $WORKFLOWS_DIR"
+  fi
+
+  # 3. write personas must consume the manifest identity at runtime
+  if wf="$(runtime_for "$id")"; then
+    [ -f "$wf" ] || fail "$id: declared runtime workflow '$wf' does not exist"
+    if ! grep -qE '(bash|sh)[[:space:]]+.*resolve-persona-identity\.sh' "$wf"; then
+      fail "$id: runtime workflow '$wf' must resolve identity from the manifest (call resolve-persona-identity.sh), not vars.BOT_USER"
+    fi
+  fi
+
+  echo "  ok: $id acts as '$account' via secret '$credential'"
+  count=$((count + 1))
+done
+
+if [ "$count" -eq 0 ]; then
+  echo "no persona manifests under $ROOT — nothing to verify."
+else
+  echo "verify-persona-identity: $count persona identity mapping(s) consistent."
+fi
diff --git a/tests/test_validate_personas.bats b/tests/test_validate_personas.bats
index 0beb11e88..e7277901f 100644
--- a/tests/test_validate_personas.bats
+++ b/tests/test_validate_personas.bats
@@ -53,6 +53,10 @@ triggers:
   surfaces: []
 trust:
   author_association_floor: [OWNER]
+runtime:
+  identity:
+    account: demo
+    credential: GH_PAT_DEMO
 evals:
   required_before: stable
   path: evals/demo/
@@ -132,6 +136,10 @@ triggers:
   surfaces: []
 trust:
   author_association_floor: [OWNER]
+runtime:
+  identity:
+    account: demo
+    credential: GH_PAT_DEMO
 evals:
   required_before: stable
   path: evals/demo/
@@ -295,3 +303,37 @@ clone_persona() {
   [ "$status" -ne 0 ]  # only 2 real < 3
   [[ "$output" == *"2 held-out"* ]]
 }
+
+@test "validate-personas REQUIRES runtime.identity" {
+  # Strip the runtime.identity block -> every persona must declare who it acts as.
+  sed '/^runtime:$/,/^    credential:/d' "$TMP/personas/demo/persona.yml" > "$TMP/personas/demo/persona.yml.tmp" && mv "$TMP/personas/demo/persona.yml.tmp" "$TMP/personas/demo/persona.yml"
+  run python3 "$VALIDATOR" "$TMP/personas" --schema "$TMP/schema.json"
+  [ "$status" -ne 0 ]
+  [[ "$output" == *"missing runtime.identity"* ]]
+}
+
+@test "validate-personas enforces the GH_PAT_<ACCOUNT> credential convention" {
+  # A non-grandfathered credential that isn't GH_PAT_* is rejected.
+  sed 's/^    credential: GH_PAT_DEMO/    credential: DEMO_TOKEN/' "$TMP/personas/demo/persona.yml" > "$TMP/personas/demo/persona.yml.tmp" && mv "$TMP/personas/demo/persona.yml.tmp" "$TMP/personas/demo/persona.yml"
+  run python3 "$VALIDATOR" "$TMP/personas" --schema "$TMP/schema.json"
+  [ "$status" -ne 0 ]
+  [[ "$output" == *"GH_PAT_<ACCOUNT>"* ]]
+}
+
+@test "validate-personas rejects identity missing the 'credential' key" {
+  # identity present but credential key omitted -> defensive check catches it cleanly.
+  sed '/^    credential:/d' "$TMP/personas/demo/persona.yml" > "$TMP/personas/demo/persona.yml.tmp"
+  mv "$TMP/personas/demo/persona.yml.tmp" "$TMP/personas/demo/persona.yml"
+  run python3 "$VALIDATOR" "$TMP/personas" --schema "$TMP/schema.json"
+  [ "$status" -ne 0 ]
+  [[ "$output" == *"runtime.identity must be a mapping"* ]]
+}
+
+@test "validate-personas rejects credential that does not match the declared account name" {
+  # GH_PAT_OTHER starts with GH_PAT_ but account is 'demo' — expected GH_PAT_DEMO.
+  sed 's/^    credential: GH_PAT_DEMO/    credential: GH_PAT_OTHER/' "$TMP/personas/demo/persona.yml" > "$TMP/personas/demo/persona.yml.tmp"
+  mv "$TMP/personas/demo/persona.yml.tmp" "$TMP/personas/demo/persona.yml"
+  run python3 "$VALIDATOR" "$TMP/personas" --schema "$TMP/schema.json"
+  [ "$status" -ne 0 ]
+  [[ "$output" == *"GH_PAT_<ACCOUNT>"* ]]
+}