Agents get read access on day one. The engineering is deciding when they get to write.
STACK_01
Agentic AI, in production
Getting a model to answer is the easy bit. Once it can act, every tool needs a scope and something that can say no.
NOW Multi-agent ops assistant on Bedrock AgentCore: a supervisor routes work to three agents (runbook RAG, ITSM, Kubernetes) through an MCP gateway on read-only credentials. Every answer passes a grounding check, every run is traced in OpenTelemetry.NOW Incident investigation agent: on a page it pulls CloudWatch metrics and logs, X-Ray traces and recent deploys, then writes a root-cause hypothesis with evidence links into the ticket. Rollbacks wait for the on-call engineer.NOW Claims intake on Bedrock Data Automation: S3 upload, EventBridge, one blueprint per document type. Low-confidence fields go to human review before anything is written downstream.
the try panels below are working versions of these stacks, on a small model that runs on your device. optional, and nothing downloads until you press load.
tryclaims intakelocal model
tryincident investigationlocal model
tryan agent on this site's toolstool calls · otel trace · local model
~/pocls -l[11]
current, earlier and own-time builds
now
pii gatewayveil
llama.cpp on the box swaps names, addresses and keys for tokens before a prompt reaches a cloud model. if it can't mask, nothing is sent
now
ci triagesift
regex cuts a CI log to the lines around the failure, the model sorts it and names the likely commit. drafts an issue, never merges
now
drift to pranchor
terraform plan finds console drift, CloudTrail says who and when, the agent opens a PR to codify or revert it. never runs apply
done
dotnet8-port
.NET Framework to .NET 8 with AWS Transform. the agent did the bulk rewrite, tests and a human review decided what merged
done
ops-text-to-sql
questions to SQL: schema retrieved first, read-only replica, row limit, query shown next to the result
done
itsm-triage
ServiceNow incidents matched to SOPs, fixes through Flow Designer and a MID Server, nothing runs until an engineer approves
done
zero-egress-rag
llama.cpp and a local vector index on one box, for data that cannot leave the network
# own time. own models, own endpoints, public challenges
lab
fine-tunequill
QLoRA on Qwen2.5-1.5B with Unsloth for ticket categories and tool-call JSON. scored against the base model, served as GGUF on llama.cpp
lab
jailbreak labpry
garak probes and PyRIT multi-turn attacks on my own models and guardrails. anything that gets through becomes a test case
lab
guardrail benchtripwire
Llama Guard and Bedrock Guardrails scored on public jailbreak and benign sets: misses and false refusals
lab
injection ctfflagrun
public prompt-injection challenges, and bounty work only inside a published scope
STACK_02
The model layer
If retrieval is wrong, the answer is wrong, whatever the model. Most of my RAG time goes into chunking, reranking and checking answers against sources, not prompts.