feat: add log sanitization for credentials, PII, and user content

.env.example

@@ -1,45 +1,140 @@
-# Agent Server Configuration
-AGENT_HOST=0.0.0.0
-AGENT_PORT=5002
-
-# LLM: default Gemini + GOOGLE_APPLICATION_CREDENTIALS_CONTENT below (USE_OPENAI_COMPAT_LLM=false or unset).
-# For RamaLama / Ollama / vLLM set USE_OPENAI_COMPAT_LLM=true and URL (and optional model/key):
-# USE_OPENAI_COMPAT_LLM=true
-# OPENAI_COMPAT_BASE_URL=http://127.0.0.1:8080/v1
-# OPENAI_COMPAT_MODEL=local
-# OPENAI_COMPAT_API_KEY=not-needed
-# AGENT_SSL_KEYFILE=/path_to/ssl_key.pem
-# AGENT_SSL_CERTFILE=/path_to/ssl_cert.pem
-
-# Python Logging
-PYTHON_LOG_LEVEL=INFO
+# ==============================================================================
+# Environment Variables
+#
+# Only secrets and infrastructure endpoints belong here.
+# All operational config (cache, middleware, filesystem, providers) lives in
+# config/agent/runtime/agent.yaml — the single source of truth.
+#
+# OpenShift: secrets come via Secrets, infra via ConfigMaps.
+# ==============================================================================
+
+# --- Environment ---
+# Set to "production" to enforce security hardening:
+# - ENABLE_AUTH must be true
+# - MCP ssl_verify cannot be disabled
+# - PII is scrubbed from error responses
+# - Security headers are enforced
+ENVIRONMENT=development
+
+# --- Security ---
+# Request body size limit (bytes) - prevents DoS attacks
+REQUEST_BODY_MAX_SIZE=10485760  # 10MB
+
+# --- SSO / OIDC Authentication ---
+# Supports any OIDC-compliant provider (Keycloak, Okta, Azure AD, Auth0, etc.)
+ENABLE_AUTH=false
+SSO_ISSUER_URL=https://sso.example.com/realms/myrealm
+SSO_CLIENT_ID=your-client-id
+SSO_CLIENT_SECRET=your-client-secret
+# SSO_JWKS_URI=https://sso.example.com/realms/myrealm/protocol/openid-connect/certs
+
+# Dev fallback identity (used when ENABLE_AUTH=false)
+SSO_DEV_USERNAME=John Doe
+SSO_DEV_USER_ID=dev-user
 
-USE_INMEMORY_SAVER=true
+# User ID encryption for observability privacy
+ENABLE_USER_ID_ENCRYPTION=false
+# USER_ID_ENCRYPTION_KEY=your-32-byte-hex-key
 
-# pgvector credentials for agentic memory (used when USE_INMEMORY_SAVER=false)
-POSTGRES_USER=pgvector
-POSTGRES_PASSWORD=pgvector
-POSTGRES_HOST=0.0.0.0
+# MCP OAuth token encryption (Fernet key — required when using auth_mode oauth/dcr)
+# Encrypts access/refresh tokens in Redis and DCR client secrets in Postgres.
+# Generate: python -c "from cryptography.fernet import Fernet; print(Fernet.generate_key().decode())"
+# MCP_TOKEN_ENCRYPTION_KEY=
+# Optional previous key during rotation (decrypt only — see README)
+# MCP_TOKEN_ENCRYPTION_KEY_PREVIOUS=
+# AGENT_PUBLIC_BASE_URL=http://localhost:5002
+
+# --- Infrastructure ---
+
+# Postgres (checkpoints, memory, feedback)
+# Local dev (`make local`): localhost + port published by compose pgvector (5432)
+POSTGRES_HOST=localhost
 POSTGRES_PORT=5432
-POSTGRES_DB=pgvector
+POSTGRES_DB=template_agent
+POSTGRES_USER=postgres
+POSTGRES_PASSWORD=postgres
+
+# Redis (Aegra broker: SSE streaming, job queue, crash recovery)
+# Local dev (`make local`): localhost + port published by compose redis (6379)
+REDIS_URL=redis://localhost:6379/0
+REDIS_BROKER_ENABLED=true
+
+# MongoDB (platform token usage rollup — optional, set by deploy components)
+#
+# SECURITY: MONGODB_URI may contain credentials in the URI itself:
+#   mongodb://user:password@host:27017/tokenusage?authSource=tokenusage
+#
+# - NEVER commit a URI with credentials to version control.
+# - NEVER log or expose this value in error messages or debug output.
+# - In production, inject via secrets management:
+#     Kubernetes : mount as a Secret, reference via envFrom or env.valueFrom.secretKeyRef
+#     AWS        : use Secrets Manager or SSM Parameter Store with an operator/init container
+#     GCP        : use Secret Manager with Workload Identity
+#     Vault      : use the Vault Agent injector or ESO (External Secrets Operator)
+# - Scope the MongoDB user to read/write on the tokenusage DB only — no admin privileges.
+# - Rotate credentials without redeploying by updating the secret and triggering a rollout.
+#
+# Local dev (unauthenticated, never in production):
+# MONGODB_URI=mongodb://localhost:27017
+# MONGODB_DB=tokenusage
+
+# --- Observability ---
 
-# exception
-LANGFUSE_SECRET_KEY=sk-lf-f46b492e-9335-
-LANGFUSE_PUBLIC_KEY=pk-lf-dfa0dab0-c486-
+# Langfuse (v4 SDK — auto-read by client and CallbackHandler)
+LANGFUSE_PUBLIC_KEY=pk-lf-...
+LANGFUSE_SECRET_KEY=sk-lf-...
 LANGFUSE_BASE_URL=https://cloud.langfuse.com
 LANGFUSE_TRACING_ENVIRONMENT=development
 
-# Google Vertex AI service creds
+# OpenTelemetry — token budget export (metrics/traces via otel_setup.py)
+# Agent lifecycle metrics (conversations, streams, threads) via observability.yaml
+# ENABLE_OTEL_METRICS=false
+# OTEL_EXPORTER_OTLP_ENDPOINT=
+# ENABLE_OTEL_TRACES=false
+# OTEL_EXPORTER_OTLP_TRACES_ENDPOINT=http://localhost:4317
+# OTEL_SERVICE_NAME=template-agent
+# OTEL_AUTH_TOKEN=
+# OTEL_METRIC_EXPORT_INTERVAL_MILLIS=10000
+# ENABLE_OTEL=false
+# OTEL_EXPORTER_OTLP_INSECURE=true
+# OTEL_METRIC_EXPORT_INTERVAL=5000
+
+# --- Model Provider Credentials ---
+
+# Google Vertex AI
 GOOGLE_APPLICATION_CREDENTIALS_CONTENT='{
     "type": "service_account",
-    "project_id": "data-and-ai-gemini",
-    ...
-    ...
+    "project_id": "your-project-id",
     ...
     "universe_domain": "googleapis.com"
   }'
 
-# MCP config
-MCP_SERVER_NAME=template-mcp-server
-MCP_SERVER_URL=http://localhost:5001/mcp
-MCP_TRANSPORT_PROTOCOL=streamable_http
+# vLLM / OpenAI-compatible (optional)
+# VLLM_BASE_URL=http://vllm-server:8000/v1
+# VLLM_API_KEY=EMPTY
+
+# --- Granite Guardian Guardrails ---
+#
+# IBM Granite Guardian provides input/output content safety checks.
+# Requires access to a running Granite Guardian model endpoint (vLLM or OpenAI-compatible).
+#
+# GUARDIAN_API_BASE=http://guardian-server:8000/v1          # Endpoint URL — guardrails activate when this is set
+# GUARDIAN_API_KEY=EMPTY                                    # API key (use EMPTY for unauthenticated vLLM)
+# GUARDIAN_SSL_VERIFY=true                                  # Set false to skip TLS verification (dev only)
+# Model is configured in config/agent/runtime/guardrails.yaml
+
+# --- Runtime (rarely changed) ---
+
+PYTHON_LOG_LEVEL=INFO
+
+# Log sanitization — redacts credentials (bearer tokens, API keys, AWS/GitHub
+# tokens) and sensitive headers from log output. Personal PII is delegated to
+# the PII middleware configured in agent.yaml.
+# LOG_SANITIZATION_ENABLED=true
+# LOG_SANITIZATION_CUSTOM_PATTERNS=INTERNAL-[0-9]{6},ACCT[0-9]+   # Comma-separated extra regexes
+# Prompts, messages and model output are logged as "<redacted: N chars>".
+# Set false only for trusted local debugging — it logs raw user input.
+# LOG_REDACT_USER_CONTENT=true
+
+# Per-MCP OAuth client secret (auth_mode: oauth — referenced via oauth.client_secret_env in mcp.json)
+# MY_OAUTH_MCP_CLIENT_SECRET=your-client-secret

.github/CODEOWNERS

@@ -0,0 +1,7 @@
+# Default owners for everything
+* @redhat-data-and-ai/template-agent-maintainers
+
+# CI and infrastructure
+.github/ @redhat-data-and-ai/template-agent-maintainers
+Containerfile @redhat-data-and-ai/template-agent-maintainers
+deployment/ @redhat-data-and-ai/template-agent-maintainers

.github/ISSUE_TEMPLATE/bug_report.yml

@@ -0,0 +1,68 @@
+name: Bug Report
+description: Report a bug or unexpected behavior
+labels: ["bug", "triage"]
+body:
+  - type: markdown
+    attributes:
+      value: |
+        Thanks for reporting a bug. Please fill in the details below so we can reproduce and fix it.
+
+  - type: textarea
+    id: description
+    attributes:
+      label: Description
+      description: A clear description of the bug.
+      placeholder: What happened?
+    validations:
+      required: true
+
+  - type: textarea
+    id: steps
+    attributes:
+      label: Steps to reproduce
+      description: Minimal steps to trigger the issue.
+      placeholder: |
+        1. Run `make local`
+        2. Send a request to ...
+        3. Observe ...
+    validations:
+      required: true
+
+  - type: textarea
+    id: expected
+    attributes:
+      label: Expected behavior
+      description: What should have happened instead?
+    validations:
+      required: true
+
+  - type: input
+    id: version
+    attributes:
+      label: Version
+      description: Output of `git describe --tags` or the container image tag.
+      placeholder: v1.2.3 or deep-agent-abc1234
+    validations:
+      required: false
+
+  - type: dropdown
+    id: environment
+    attributes:
+      label: Environment
+      options:
+        - Local (make local)
+        - Container (make container)
+        - Kind (make kind)
+        - OpenShift
+        - Other
+    validations:
+      required: true
+
+  - type: textarea
+    id: logs
+    attributes:
+      label: Logs or error output
+      description: Paste relevant logs (redact any secrets).
+      render: shell
+    validations:
+      required: false

.github/ISSUE_TEMPLATE/config.yml

@@ -0,0 +1,5 @@
+blank_issues_enabled: false
+contact_links:
+  - name: Security vulnerability
+    url: https://github.com/redhat-data-and-ai/template-agent/security/advisories/new
+    about: Report security issues privately via GitHub Security Advisories.

.github/ISSUE_TEMPLATE/feature_request.yml

@@ -0,0 +1,41 @@
+name: Feature Request
+description: Suggest a new feature or improvement
+labels: ["enhancement"]
+body:
+  - type: markdown
+    attributes:
+      value: |
+        Describe the feature you'd like and the problem it solves.
+
+  - type: textarea
+    id: problem
+    attributes:
+      label: Problem
+      description: What problem does this solve? What's the current pain point?
+      placeholder: I'm always frustrated when ...
+    validations:
+      required: true
+
+  - type: textarea
+    id: solution
+    attributes:
+      label: Proposed solution
+      description: How should this work?
+    validations:
+      required: true
+
+  - type: textarea
+    id: alternatives
+    attributes:
+      label: Alternatives considered
+      description: Other approaches you've thought about.
+    validations:
+      required: false
+
+  - type: textarea
+    id: context
+    attributes:
+      label: Additional context
+      description: Links, screenshots, or related issues.
+    validations:
+      required: false

.github/PULL_REQUEST_TEMPLATE.md

@@ -0,0 +1,28 @@
+## What
+
+<!-- One or two sentences: what changed and why. Link the issue. -->
+
+Fixes #
+
+## How
+
+<!-- Describe your approach. Call out non-obvious design decisions, trade-offs, or alternatives you considered. -->
+
+## Testing
+
+<!-- How did you verify this works? -->
+
+- [ ] Unit tests added/updated
+- [ ] Ran locally (`uv run pytest tests/unit -x`)
+- [ ] Manual verification (describe below if applicable)
+
+## Rollback
+
+<!-- How would you revert this if it breaks production? "Revert the commit" is fine for most changes. -->
+
+## Checklist
+
+- [ ] PR title follows [Conventional Commits](https://www.conventionalcommits.org/) (`feat:`, `fix:`, `ci:`, etc.)
+- [ ] No secrets, credentials, or PII in the diff
+- [ ] No breaking changes (or documented above with a migration path)
+- [ ] Pre-commit hooks pass (`uv run pre-commit run --all-files`)

.github/actions/setup-python-uv/action.yml

@@ -0,0 +1,26 @@
+name: Setup Python with uv
+description: Install uv, create a virtual environment, and install project dependencies
+
+inputs:
+  python-version:
+    description: Python version to use
+    required: false
+    default: "3.14"
+
+runs:
+  using: composite
+  steps:
+    - name: Install uv
+      uses: astral-sh/setup-uv@caf0cab7a618c569241d31dcd442f54681755d39 # v3
+      with:
+        version: "latest"
+        enable-cache: true
+        cache-dependency-glob: "**/pyproject.toml"
+
+    - name: Set up virtual env
+      shell: bash
+      run: uv venv --python ${{ inputs.python-version }}
+
+    - name: Install dependencies
+      shell: bash
+      run: uv pip install -e ".[dev]"

.github/dependabot.yml

@@ -0,0 +1,26 @@
+version: 2
+updates:
+  - package-ecosystem: pip
+    directory: "/"
+    schedule:
+      interval: weekly
+    groups:
+      minor-and-patch:
+        update-types:
+          - minor
+          - patch
+
+  - package-ecosystem: docker
+    directory: "/"
+    schedule:
+      interval: weekly
+
+  - package-ecosystem: github-actions
+    directory: "/"
+    schedule:
+      interval: weekly
+    groups:
+      actions-minor-patch:
+        update-types:
+          - minor
+          - patch