https://koalstingkdelaney-gif.github.io/kestrelattice/articles/ Agent governance playbooks, checklists, and field notes from ghostcorpnet. en https://koalstingkdelaney-gif.github.io/kestrelattice/articles/ai-agent-vendor-risk-assessment.html https://koalstingkdelaney-gif.github.io/kestrelattice/articles/ai-agent-vendor-risk-assessment.html Thu, 01 Oct 2026 00:00:00 GMT 20 due-diligence questions to ask before buying a third-party AI agent: identity, approvals, spend controls, audit logs, data handling, exit terms. https://koalstingkdelaney-gif.github.io/kestrelattice/articles/ai-agent-incident-response-plan.html https://koalstingkdelaney-gif.github.io/kestrelattice/articles/ai-agent-incident-response-plan.html Thu, 01 Oct 2026 00:00:00 GMT Build a robust AI agent incident response plan to stop runaway loops, prompt injections, and unauthorized actions before they break production. https://koalstingkdelaney-gif.github.io/kestrelattice/articles/ai-agent-audit-log-schema-retention.html https://koalstingkdelaney-gif.github.io/kestrelattice/articles/ai-agent-audit-log-schema-retention.html Thu, 01 Oct 2026 00:00:00 GMT What belongs in an AI agent audit log: the minimum schema, tamper-evident storage, retention tiers, and what reviewers actually ask for. https://koalstingkdelaney-gif.github.io/kestrelattice/articles/ai-agent-policy-templates-what-to-include.html https://koalstingkdelaney-gif.github.io/kestrelattice/articles/ai-agent-policy-templates-what-to-include.html Wed, 30 Sep 2026 00:00:00 GMT Writing AI agent policies from scratch wastes weeks. Here is exactly what a complete agent policy pack contains — and the ready-made templates that cover each piece. https://koalstingkdelaney-gif.github.io/kestrelattice/articles/us-state-ai-laws-agents-2026.html https://koalstingkdelaney-gif.github.io/kestrelattice/articles/us-state-ai-laws-agents-2026.html Wed, 30 Sep 2026 00:00:00 GMT Texas TRAIGA, California's AI transparency laws, and shifting Colorado rules — the 2026 US state AI law landscape for teams running AI agents, and the practical playbook. https://koalstingkdelaney-gif.github.io/kestrelattice/articles/eu-ai-act-agent-evidence-checklist.html https://koalstingkdelaney-gif.github.io/kestrelattice/articles/eu-ai-act-agent-evidence-checklist.html Wed, 30 Sep 2026 00:00:00 GMT EU AI Act enforcement began August 2, 2026. If your company deploys AI agents that touch EU users, here is the evidence checklist regulators and enterprise customers expect. https://koalstingkdelaney-gif.github.io/kestrelattice/articles/ai-agent-permissions-audit.html https://koalstingkdelaney-gif.github.io/kestrelattice/articles/ai-agent-permissions-audit.html Wed, 30 Sep 2026 00:00:00 GMT A practical permissions audit for AI agents in production: 12 questions covering grants, credential handling, spend ceilings, and audit trails — before something breaks. https://koalstingkdelaney-gif.github.io/kestrelattice/articles/runaway-agent-costs-spend-ceilings.html https://koalstingkdelaney-gif.github.io/kestrelattice/articles/runaway-agent-costs-spend-ceilings.html Wed, 30 Sep 2026 00:00:00 GMT A dev watched an AI agent burn $700+ in 72 hours. The fix is not a better dashboard — it is per-plan spend ceilings that fail closed, plus three more controls that cost $0. https://koalstingkdelaney-gif.github.io/kestrelattice/articles/selling-ai-products-to-europe.html https://koalstingkdelaney-gif.github.io/kestrelattice/articles/selling-ai-products-to-europe.html Wed, 30 Sep 2026 00:00:00 GMT EU AI Act enforcement is pushing audit-trail requirements into B2B procurement. How AI vendors can build a governance evidence pack that shortens enterprise sales cycles. https://koalstingkdelaney-gif.github.io/kestrelattice/articles/scoped-expiring-grants-agents.html https://koalstingkdelaney-gif.github.io/kestrelattice/articles/scoped-expiring-grants-agents.html Wed, 30 Sep 2026 00:00:00 GMT Never hand an agent your raw production key. Issue a grant instead — a record with five explicit fields, three rules, and no auto-grant path, ever. https://koalstingkdelaney-gif.github.io/kestrelattice/articles/observability-vs-governance.html https://koalstingkdelaney-gif.github.io/kestrelattice/articles/observability-vs-governance.html Wed, 30 Sep 2026 00:00:00 GMT Teams buy observability and call it governance. But dashboards show what happened; only fail-closed controls stop what should not happen. The $0 spend ceiling that does it. https://koalstingkdelaney-gif.github.io/kestrelattice/articles/ley-ia-ue-lista-verificacion-agentes.html https://koalstingkdelaney-gif.github.io/kestrelattice/articles/ley-ia-ue-lista-verificacion-agentes.html Wed, 30 Sep 2026 00:00:00 GMT La Ley de IA de la UE entró en aplicación el 2 de agosto de 2026. Si tu empresa usa agentes de IA que llegan a usuarios europeos, esta es tu lista de verificación práctica. https://koalstingkdelaney-gif.github.io/kestrelattice/articles/governance-economics-paying-wrong-layer.html https://koalstingkdelaney-gif.github.io/kestrelattice/articles/governance-economics-paying-wrong-layer.html Wed, 30 Sep 2026 00:00:00 GMT LangSmith is $39/seat/month. AgentOps runs $49–$199/month. Both sell observability. The layer that actually stops incidents — policy and enforcement — costs $0 to own. https://koalstingkdelaney-gif.github.io/kestrelattice/articles/buy-observability-build-governance.html https://koalstingkdelaney-gif.github.io/kestrelattice/articles/buy-observability-build-governance.html Wed, 30 Sep 2026 00:00:00 GMT Prompt engineering cannot enforce boundaries on multi-agent systems. Score vendors against the governance gap — then build the kill switch yourself. https://koalstingkdelaney-gif.github.io/kestrelattice/articles/agent-permissions-tier-by-consequence.html https://koalstingkdelaney-gif.github.io/kestrelattice/articles/agent-permissions-tier-by-consequence.html Wed, 30 Sep 2026 00:00:00 GMT Who can use which agent? What can each agent access? A three-tier permission model — hard-gated, bounded auto, draft-only — plus the heuristics for tiering reads with write-grade consequences. https://koalstingkdelaney-gif.github.io/kestrelattice/articles/agent-governance-faq.html https://koalstingkdelaney-gif.github.io/kestrelattice/articles/agent-governance-faq.html Wed, 30 Sep 2026 00:00:00 GMT Straight answers to the most-asked agent governance questions from practitioner communities: prompt fatigue, runaway costs, agent inventory, and payment permissions.