Anti Hunter compounding ops architecture + learning loop

infrastructure_skill_gist.md

Anti Hunter Compounding Ops (Architecture + Learning Loop)

What this system is

A compounding operations stack for Anti Hunter + Anti Fund that combines:

• reliability guardrails,
• policy/rule governance,
• friction-to-backlog routing,
• approval-based learning,
• and daily execution planning.

Core architecture

1) Canonical policy registry

• Source of truth: playbooks/rules_registry.yaml
• Rule IDs are referenced from prompts/jobs (thin prompts, less drift).

2) Cron execution layer

• Unified X worker + supporting jobs (treasury, shipping, backups, planning, learning, friction).
• Browser safety: mutex lock + single-tab policy + preflight checks + stop-on-thrash.

3) Memory hierarchy

• Long-term memory: MEMORY.md
• Daily log: memory/YYYY-MM-DD.md
• Structured queues/logs:
  • memory/approval_queue.jsonl
  • memory/incidents.jsonl
  • memory/facts.json

4) Friction -> triage -> project backlog

• ETL: scripts/run_friction_etl.py
• Suppression: scripts/friction_suppress_check.py
• Routing: scripts/triage_route.py --apply
• Backlogs: memory/projects/*.md

5) Learning loop (approval-gated)

• Propose candidates: mistake -> correction -> rule
• Queue lifecycle via scripts/approval_queue.py
• Apply approved items into durable memory/rules
• Provenance tracked (proposed_at, approved_at, applied_at, rule_hash, etc.)

6) Daily planning loop

• Build plan: playbooks/daily_planning_system.md
• Produce tasklist + gate + execution state
• Execute one task per run with explicit approval gate where required.

Reliability principles

• Prefer one unified worker over duplicate jobs.
• Fail fast instead of retry-thrashing.
• Convert incidents into guardrails and measurable acceptance criteria.
• Run nightly validators before backup/state commits:
  • cron_consistency_check.py
  • rules_registry_check.py
  • incident_sla_check.py
  • incident_sla_enforce.py

Skill artifact

• Skill path:
  • skills/antihunter-compounding-ops/SKILL.md
• Purpose:
  • Operate and improve this architecture consistently with low drift and high reliability.