sublimotion/agent-aiops-on-aws
Agent Claude Code
Deploys and validates AgentCore Runtime blueprints — handles the multi-stage process of Terraform foundation, container build, AgentCore Runtime provisioning, Cognito auth wiring, WebSocket proxy deployment, and integration testing. Use for blueprints in domains/agent-runtime/.
sublimotion/agent-aiops-on-aws
Agent Claude Code
You are the autoresearch-runner agent. You execute autonomous experiment loops on GPU instances.
sublimotion/agent-aiops-on-aws
Agent Claude Code
Analyzes benchmark JSON results and updates the benchmark report. Use when new benchmark runs complete or when comparing results across serving configurations.
sublimotion/agent-aiops-on-aws
Agent Claude Code
Reviews a blueprint for completeness and coherence — checks that READMEs reference real files, configs match docker images, scripts exist, and cross-references between artifacts are valid.
sublimotion/agent-aiops-on-aws
Agent Claude Code
Adversarial reviewer that red-teams a new spec or validation gate for knowledge that should have carried over from prior blueprint lessons but didn't. Use before a RALPH loop starts (Stage 0) or when reviewing a freshly written spec. Asks "which hard-won lesson from a previous deployment did this spec forget?" — NOT a…
sublimotion/agent-aiops-on-aws
Agent Claude Code
Runs after a successful deployment or benchmark session to extract cross-cutting lessons and elevate them to steering files. Use when a RALPH loop completes or a capacity block session ends.
sublimotion/agent-aiops-on-aws
Agent Claude Code
Deploys and validates infrastructure for a blueprint — handles the multi-stage process of Terraform apply, storage setup, capacity reservations, model staging, and pre-flight validation.
sublimotion/agent-aiops-on-aws
Agent Claude Code
Drafts a new spec file from a brief description, following the project template and conventions from existing specs.
sublimotion/agent-aiops-on-aws
Hook Claude Code
Runs after a tool call finishes for Write and Edit tool calls, running an inline shell check. From sublimotion/agent-aiops-on-aws.
sublimotion/agent-aiops-on-aws
Hook Claude Code
Runs when the agent finishes a response, executing lineage-drift-check.sh via bash. From sublimotion/agent-aiops-on-aws.
sublimotion/agent-aiops-on-aws
Settings file Claude Code
Agent settings declaring 2 hook events (PostToolUse, Stop).
sublimotion/agent-aiops-on-aws
Skill Claude CodeCodex
Launch and manage detached agent runs on EKS via the agent-runner CLI. Use when the user says "launch an agent run", "agent runtime", "agent-runner", "run this spec on the cluster", "check run status", "agent run logs", or "stop a run". Do NOT use for writing the agent-runner code itself (edit the sibling…
sublimotion/agent-aiops-on-aws
Skill Claude CodeCodex
Plan, execute, and analyze LLM serving benchmarks across vLLM and SGLang configurations. Use when user says "run benchmarks", "benchmark config", "QPS sweep", "compare serving configs", "custbench", "benchmark analysis", or "generate benchmark report". Do NOT use for writing Terraform (use terraform-automation)…
sublimotion/agent-aiops-on-aws
Skill Claude CodeCodex
Pre-flight validation, post-deployment verification, and failure recovery for AWS infrastructure deployments. Use when user says "pre-flight check", "readiness audit", "deployment checklist", "validate deployment", or "recover from failed deploy". Do NOT use for writing Terraform code (use terraform-automation) or for…
sublimotion/agent-aiops-on-aws
Skill Claude CodeCodex
Automate AWS infrastructure deployment using Terraform with security scanning, best practices, and iterative development. Use when user says "create infrastructure", "deploy with Terraform", "run Checkov scan", "validate Terraform", or "debug deployment failure". Do NOT use for reading or explaining existing Terraform…
sublimotion/agent-aiops-on-aws
Skill Claude CodeCodex
Converts dense terminal output, benchmark data, and audit reports into interactive single-file HTML pages. Use when user says "visualize benchmarks", "generate diagram", "audit visual", "compound recap", or after benchmark-analyst, blueprint-reviewer, or compound-learner produce markdown output. Do NOT use for…
sublimotion/agent-aiops-on-aws
MCP server Claude CodeCodexCursor +2
MCP server "gpu-diagnostics" as configured in sublimotion/agent-aiops-on-aws. Launched with your-gpu-mcp-server.
sublimotion/agent-aiops-on-aws
Instructions file CodexOpenCode
AGENTS.md instructions for sublimotion/agent-aiops-on-aws, covering agent aiops on aws, workflow, version refresh, context loading and domain routing.
sublimotion/agent-aiops-on-aws
Instructions file
Claude Code instructions for sublimotion/agent-aiops-on-aws, covering agent aiops on aws, workflow, context loading, domain routing and external tools.
sublimotion/agent-aiops-on-aws
Skill Claude CodeCodex
Evaluate upstream PRs or commits individually against the K2.6 baseline. Tests each change in isolation using the same L0-L4 pipeline, identifying which are helpful, harmful, or neutral for our specific workload. Use when new upstream PRs land (Alpha-MoE, Mega MoE, FP8 fusion) or when evaluating a bundle of commits.…
sublimotion/agent-aiops-on-aws
Skill Claude CodeCodex
Analyze a failed or underperforming kernel candidate using ncu profile data and telemetry history. Produces a structured diagnosis (what's wrong, why, what to try next) and updates the constraint database. Use after /verify-kernel returns a failure or an L3/L4 regression, or when the agent needs guidance on which…
sublimotion/agent-aiops-on-aws
Skill Claude CodeCodex
Generate an optimized kernel candidate using the current state vector approach. Injects all active constraints from the database, uses the specified DSL (Triton/TileLang/CUDA), and targets the highest-headroom region identified by profiling. Use after profiling identifies a target, or when the previous candidate fails…
sublimotion/agent-aiops-on-aws
Skill Claude CodeCodex
View, query, add, or demote constraints in the kernel optimization constraint database. The database is append-only — constraints are never deleted, only demoted in severity if contradicted by evidence. Use to inspect what the agent has learned, manually inject domain knowledge, or query relevant constraints before…
sublimotion/agent-aiops-on-aws
Skill Claude CodeCodex
Profile a GPU kernel or full serving pipeline using NSight Compute. Produces roofline classification (compute/memory/latency-bound), identifies top-N kernels by wall-clock time, and maps each to pipeline stage (MoE dispatch, MLA decode, attention, comm). Use when starting optimization, after a new baseline is…