agent management platform - Detect agent conflicts. Track costs. Audit everything.

Hey Product Hunt! I'm Leo, the maker of Ahy Governance. The problem: Last month I deployed 5 AI agents to handle contract review. Within 3 days: - Agent A said "deadline is June 30" - Agent B said "deadline is July 15" - Nobody noticed until a client flagged it - The token bill hit $200 with zero visibility into which agent cost what I looked for tools to fix this. LangSmith and LangFuse are great — but they trace individual LLM calls. They don't understand multi-agent orchestration. They can't detect when two agents produce contradictory outputs. They can't tell you how much each agent costs. So I built Ahy Governance. 7 modules, 312 tests, MIT licensed. What makes it different: 1. Conflict Detector — 5 types of inter-agent conflicts (fact, format, dependency, scope, confidence). Currently unique in the market. 2. Cost Tracker — per-agent attribution with 22 model pricing tables and budget circuit breakers 3. Audit Logger — SHA-256 hash chain, tamper-proof. Export SOC2 or ISO27001 compliance reports with one click. 4. Health Monitor — heartbeat, P50/P95/P99 latency, pipeline DAG tracking 5. RBAC + API Keys — 3-tier roles, multi-tenant isolation 6. Prompt Guard — 13 injection patterns + PII redaction 7. Memory Sharing — namespaced key-value store shared across agents Pricing: Community (free, 1 agent) → Pro ($99/mo, 10 agents) → Team ($299/mo, 50 agents) → Enterprise ($999/mo, unlimited). Tech stack: Python >=3.10, FastAPI backend, single-file HTML dashboard. Zero npm, zero build step. `pip install ahy-governance[web] && ahy-dashboard` and you're running. I'd love feedback on: - What's missing before you'd use this in production? - Which module should I tackle next? - Would you pay for this? If not, what would make it worth paying for? Thanks for checking it out!