Agent Infrastructure Bundle

Everything your agent needs on day one.

Three production-tested frameworks that give any AI agent a complete operational foundation:

1. Agent Trust & Safety Framework (SECURITY.md)

• Core principle that blocks most prompt injection attacks
• Three-tier trust levels (Autonomous / Approval Required / Off-Limits)
• The Symmetry Test for edge-case decisions
• Spending controls with configurable thresholds
• 6 attack vectors with specific defenses
• Role-specific customization guide

2. Agent Heartbeat — Autonomous Operations (HEARTBEAT.md)

• Two-step autonomous work cycle
• Customizable 4-tier priority stack
• Session handoff protocol
• Crash recovery procedure
• Daily logging format (OUTCOME/SCORE/WHY)
• Boot sequence for consistent startup

3. Agent Memory Framework (MEMORY.md)

• Two-tier memory: Core (never expires) + Working (monthly review)
• PARA organization for topic-based retrieval
• Daily notes with structured event logging
• Source tagging for every fact (direct/inferred/derived)
• Common mistakes guide

Why a Bundle?

Security without operations = safe but idle. Operations without memory = productive but amnesiac. Memory without security = remembers everything but vulnerable.

Together, they form the minimum viable infrastructure for any production agent.

Who This Is For

• Setting up a new agent and want it production-ready immediately
• Tired of agents that forget everything, work reactively, and have no guardrails
• Want three tested frameworks that work together, not three separate shopping trips

$16 — Save 40% vs buying separately ($27). Three files. 30 minutes to customize. Complete agent infrastructure.

One-time purchase. No dependencies. Works on any agent with workspace files (OpenClaw, Claude, Codex, or custom).