Replace real PHI in fixtures with John Smith USA mock data

.githooks/pre-push

@@ -1,4 +1,4 @@
 #!/usr/bin/env bash
 set -euo pipefail
 ROOT="$(cd "$(dirname "$0")/.." && pwd)"
-exec python3 "$ROOT/scripts/check_safe_to_push.py" --public
+exec python3 "$ROOT/scripts/check_safe_to_push.py" --public --push

.github/dependabot.yml

@@ -0,0 +1,10 @@
+version: 2
+updates:
+  - package-ecosystem: github-actions
+    directory: /
+    schedule:
+      interval: weekly
+  - package-ecosystem: pip
+    directory: /
+    schedule:
+      interval: weekly

.github/workflows/secret-scan.yml

@@ -5,6 +5,9 @@ on:
     branches: [main]
   pull_request:
 
+permissions:
+  contents: read
+
 jobs:
   paths-and-secrets:
     runs-on: ubuntu-latest
@@ -20,14 +23,43 @@ jobs:
             echo "Patient corpus files must not be in the public repo"
             exit 1
           fi
-          for f in bot_config.json .env; do
+          if git ls-files '**/.env' | grep -v '\.env\.example$'; then
+            echo "Blocked .env file(s) in repo"
+            exit 1
+          fi
+          if git ls-files '**/bot_config.json' 'bot_config.json' 2>/dev/null | grep -q .; then
+            echo "bot_config.json must not be committed"
+            exit 1
+          fi
+          for f in sources incoming; do
             if git ls-files "$f" 2>/dev/null | grep -q .; then
-              echo "Blocked file in repo: $f"
+              echo "Blocked path in repo: $f"
               exit 1
             fi
           done
 
+      - name: Scan tracked tree for secrets and PHI
+        run: python3 scripts/check_safe_to_push.py --public --scan-all
+
       - name: Gitleaks
         uses: gitleaks/gitleaks-action@v2
         env:
           GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+        with:
+          config-path: .gitleaks.toml
+
+  dependency-audit:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+
+      - uses: actions/setup-python@v5
+        with:
+          python-version: "3.12"
+
+      - name: Dependency audit
+        run: |
+          python -m pip install --upgrade pip
+          pip install pip-audit
+          pip install -e .
+          pip-audit

.gitleaks.toml

@@ -0,0 +1,16 @@
+title = "biohackbot gitleaks"
+
+[allowlist]
+description = "Ignore mock patient tokens in public fixtures"
+paths = [
+  '''tests/fixtures/''',
+  '''examples/demo-instance/''',
+]
+
+[[allowlist.regexes]]
+description = "Mock US insurance id in fixtures"
+regex = '''US-MOCK-INS-'''
+
+[[allowlist.regexes]]
+description = "Mock deploy placeholders"
+regex = '''TELEGRAM_BOT_TOKEN=\s*$'''

README.md

@@ -57,7 +57,7 @@ This creates:
 Edit `PATIENT_PROFILE.json`:
 
 ```json
-{"dob": "1985-03-20", "full_name_ru": "Your Name"}
+{"dob": "1985-06-15", "full_name_ru": "John Smith", "country": "USA"}
 ```
 
 ## 3. Add documents

README.ru.md

@@ -55,7 +55,7 @@ medbots init ~/my-health
 Отредактируйте `PATIENT_PROFILE.json`:
 
 ```json
-{"dob": "1985-03-20", "full_name_ru": "Иван Иванов"}
+{"dob": "1985-06-15", "full_name_ru": "John Smith", "country": "USA"}
 ```
 
 ## 3. Добавление документов

README.zh-CN.md

@@ -57,7 +57,7 @@ medbots init ~/my-health
 编辑 `PATIENT_PROFILE.json`：
 
 ```json
-{"dob": "1985-03-20", "full_name_ru": "Your Name"}
+{"dob": "1985-06-15", "full_name_ru": "John Smith", "country": "USA"}
 ```
 
 ## 3. 添加文档

SECURITY.md

@@ -1,12 +1,39 @@
 # Security
 
+## Public repository policy
+
 **Never commit to this public repo:**
 
-- API keys, bot tokens, `.env` files
-- `bot_config.json` with real paths
+- API keys, bot tokens, real `.env` files
+- `bot_config.json` with instance-specific paths
 - Patient data: labs, genetics, clinical notes, `PATIENT_PROFILE.json`, `pdf_text/`, `doc_text/`
-- Raw PDFs in `sources/`
+- Raw PDFs in `sources/` or `incoming/`
 
 Keep health data in a local or private instance directory (see README).
 
+## Automated controls
+
+| Control | Purpose |
+|---------|---------|
+| `scripts/check_safe_to_push.py --public --push` | Pre-push hook: scans outgoing commits (not just staged files) |
+| `scripts/check_safe_to_push.py --public --scan-all` | CI: scans all tracked files |
+| `.github/workflows/secret-scan.yml` | Gitleaks + path blocks + dependency audit |
+| `.gitleaks.toml` | Allowlist for mock fixtures only |
+| `deploy/lib/validate-deploy-path.sh` | Blocks shell metacharacter injection in deploy vars |
+
+Install hooks after clone:
+
+```bash
+git config core.hooksPath .githooks
+```
+
+## Private VPS / Telegram bot
+
+- Store tokens in `~/.config/medbot-openclaw.env` with `chmod 600`
+- Set `TELEGRAM_ALLOWED_USER_IDS` — never run the bot open to all Telegram users
+- Corpus on VPS: `/opt/medbot-corpus` — restrict SSH and filesystem permissions
+- Do not expose corpus directory via HTTP without authentication
+
+## Reporting
+
 Report vulnerabilities via [GitHub Security Advisories](https://github.com/apodobe/biohackbot/security/advisories/new) — without attaching PHI or tokens.

deploy/02-rsync-corpus.sh

@@ -1,15 +1,25 @@
 #!/usr/bin/env bash
 set -euo pipefail
+SCRIPT_DIR="$(cd "$(dirname "$0")" && pwd)"
+# shellcheck source=lib/validate-deploy-path.sh
+source "$SCRIPT_DIR/lib/validate-deploy-path.sh"
+
 VPS="${VPS:?Set VPS=root@your-host}"
 LOCAL="${CORPUS:?Set CORPUS=/path/to/structured_database}"
 REMOTE_DIR="${REMOTE_DIR:-/opt/medbot-corpus/structured_database}"
+
+validate_deploy_value "VPS" "$VPS"
+validate_deploy_value "LOCAL" "$LOCAL"
+validate_deploy_value "REMOTE_DIR" "$REMOTE_DIR"
+
 echo "Local:  $LOCAL"
 echo "Remote: $VPS:$REMOTE_DIR"
-ssh "$VPS" "mkdir -p '$REMOTE_DIR'"
+ssh -- "$VPS" mkdir -p -- "$REMOTE_DIR"
 rsync -avz --delete \
   --exclude '.pytest_cache' \
   --exclude '__pycache__' \
   --exclude '*.pdf' \
   "$LOCAL/" "$VPS:$REMOTE_DIR/"
-ssh "$VPS" "mkdir -p /opt/medbot-corpus && chmod -R a+rX /opt/medbot-corpus"
+ssh -- "$VPS" mkdir -p -- /opt/medbot-corpus
+ssh -- "$VPS" chmod -R a+rX -- /opt/medbot-corpus
 echo "Done."

deploy/03-install-skill-on-vps.sh

@@ -1,10 +1,18 @@
 #!/usr/bin/env bash
 set -euo pipefail
+SCRIPT_DIR="$(cd "$(dirname "$0")" && pwd)"
+# shellcheck source=lib/validate-deploy-path.sh
+source "$SCRIPT_DIR/lib/validate-deploy-path.sh"
+
 VPS="${VPS:-n8n-server}"
 SKILL_SRC="$(cd "$(dirname "$0")" && pwd)/skills/biohacking-corpus"
 REMOTE_SKILLS="${REMOTE_SKILLS:-/root/.openclaw/workspace/skills}"
+
+validate_deploy_value "VPS" "$VPS"
+validate_deploy_value "REMOTE_SKILLS" "$REMOTE_SKILLS"
+
 echo "Install skill: $SKILL_SRC -> $VPS:$REMOTE_SKILLS/biohacking-corpus/"
-ssh "$VPS" "mkdir -p '$REMOTE_SKILLS/biohacking-corpus'"
+ssh -- "$VPS" mkdir -p -- "$REMOTE_SKILLS/biohacking-corpus"
 rsync -avz "$SKILL_SRC/" "$VPS:$REMOTE_SKILLS/biohacking-corpus/"
-ssh "$VPS" "chmod +x '$REMOTE_SKILLS/biohacking-corpus/'*.sh 2>/dev/null || true"
+ssh -- "$VPS" chmod +x -- "$REMOTE_SKILLS/biohacking-corpus/"*.sh 2>/dev/null || true
 echo "Done. Restart OpenClaw gateway if needed."

deploy/RUNBOOK.md

@@ -31,7 +31,8 @@ Copy and fill env template on the server:
 
 ```bash
 cp deploy/openclaw.env.example ~/.config/medbot-openclaw.env
-# set TELEGRAM token and allowed user IDs
+chmod 600 ~/.config/medbot-openclaw.env
+# set TELEGRAM token and allowed user IDs (never leave ALLOWED_USER_IDS empty in production)
 ```
 
 Skill reads corpus from `/opt/medbot-corpus/structured_database` (adjust in `02-rsync-corpus.sh` if needed).

deploy/lib/validate-deploy-path.sh

@@ -0,0 +1,22 @@
+#!/usr/bin/env bash
+# Reject shell metacharacters in deploy paths/hosts (SSH injection guard).
+validate_deploy_value() {
+  local label="$1"
+  local value="$2"
+  if [[ -z "$value" ]]; then
+    echo "ERROR: $label is empty" >&2
+    return 1
+  fi
+  if [[ "$value" == *$'\n'* ]] || [[ "$value" == *$'\r'* ]]; then
+    echo "ERROR: $label contains newline" >&2
+    return 1
+  fi
+  if [[ ! "$value" =~ ^[a-zA-Z0-9@._:/-]+$ ]]; then
+    echo "ERROR: $label contains unsafe characters: $value" >&2
+    return 1
+  fi
+  if [[ "$value" == *".."* ]]; then
+    echo "ERROR: $label must not contain .." >&2
+    return 1
+  fi
+}

deploy/openclaw.env.example

@@ -1,4 +1,5 @@
 # /root/.config/medbot-openclaw.env on VPS — do not commit real tokens
+# chmod 600 this file after filling values
 TELEGRAM_BOT_TOKEN=
 TELEGRAM_ALLOWED_USER_IDS=123456789
 MEDBOTS_CORPUS_PATH=/opt/medbot-corpus/structured_database

deploy/skills/biohacking-corpus/biohacking-corpus-path.sh

@@ -1,3 +1,8 @@
 #!/usr/bin/env bash
 # Default corpus path for biohacking-corpus skill
-echo "${BIOHACKING_CORPUS_PATH:-/opt/medbot-corpus/structured_database}"
+BASE="${BIOHACKING_CORPUS_PATH:-/opt/medbot-corpus/structured_database}"
+if [[ "$BASE" == *$'\n'* ]] || [[ ! "$BASE" =~ ^[a-zA-Z0-9/_.-]+$ ]] || [[ "$BASE" == *".."* ]]; then
+  echo "ERROR: invalid BIOHACKING_CORPUS_PATH" >&2
+  exit 1
+fi
+echo "$BASE"

docs/PARSERS.md

@@ -42,7 +42,7 @@ medbots validate-apple-health --corpus ~/my-health/structured_database
 2. **Set date of birth** in `structured_database/PATIENT_PROFILE.json` — used for age-based reference ranges and Apple Health quality checks:
 
    ```json
-   {"dob": "1985-03-20", "full_name_ru": "Your Name"}
+   {"dob": "1985-06-15", "full_name_ru": "John Smith", "country": "USA"}
    ```
 
 3. **Copy source files** into the right folder (see vendor sections below).

docs/PARSERS.ru.md

@@ -42,7 +42,7 @@ medbots validate-apple-health --corpus ~/my-health/structured_database
 2. **Указать дату рождения** в `structured_database/PATIENT_PROFILE.json` — нужна для референсов и проверки Apple Health:
 
    ```json
-   {"dob": "1985-03-20", "full_name_ru": "Иван Иванов"}
+   {"dob": "1985-06-15", "full_name_ru": "John Smith", "country": "USA"}
    ```
 
 3. **Положить файлы** в нужную папку `sources/` (см. разделы по вендорам).

examples/demo-instance/README.md

@@ -1,7 +1,9 @@
-# Demo instance (synthetic, no PHI)
+# Demo instance (synthetic mock patient)
 
 Pre-built manifest + `pdf_text/` for three vendors. **No PDF files required.**
 
+Mock patient: **John Smith**, USA (`PATIENT_PROFILE.json`).
+
 ```bash
 # from repo root
 medbots structure --bot-root examples/demo-instance
@@ -16,4 +18,4 @@ medbots pipeline --bot-root examples/demo-instance
 
 Medsi entry includes `user_drop_batch: true` (required by the Medsi parser gate in `local_structure_pdfs`).
 
-Text fixtures are copied from `tests/fixtures/pdf_text/` (redacted synthetic extracts).
+Text fixtures are copied from `tests/fixtures/pdf_text/` (synthetic extracts with mock PII).

examples/demo-instance/structured_database/PATIENT_PROFILE.json

@@ -1 +1 @@
-{"dob":"1985-03-20","full_name_ru":"Demo Patient"}
+{"dob": "1985-06-15", "full_name_ru": "John Smith", "country": "USA"}

examples/demo-instance/structured_database/pdf_text/sources__emias__2026-06-17__drop_Биохимический_анализ_крови__929313a8.pdf.txt

@@ -1,20 +1,20 @@
 Биохимический анализ крови
-Тестов
+Smith
 Фамилия: 
 № направления: 
-268953196
+MOCK-REF-90001
 Дата: 16.06.2026
-Алексей
+John
 Имя: 
 ЛПУ:
-Карта: 2364501
-Юрьевич
+Карта: MOCK-CARD-10001
+-
 Отчество: 
 Отделение:
 10.01. Терапевтическое отделение
-Врач: Михайлова Юлия Игоревна
+Врач: Robert Miller, MD
 Номер заказа:
-1011274602
+MOCK-ORDER-80001
 Наименование исследования
 Результат
 Ед. изм.
@@ -169,6 +169,6 @@ Consensus Panel. Eur Heart J.2016;37:1944-1958
 
 
 
-Калачева Ольга Сергеевна
+Lab Tech Demo
 16.06.2026 12:22:13
 Врач КДЛ: