Matthieu Vlad
Community@using-system · Paris, France
Devops & Cloud Architect
Agent Skills by Matthieu Vlad
Showing 9 vetted skills indexed across 1 GitHub repositories.
using-systemotel-guides
Maps official OpenTelemetry documentation by language for planning instrumentation of traces, metrics, logs, and profiles.
using-systemodd-memory
Persist, recall, and show committed observability memory artifacts under .odd/ in git repositories.
using-systemobservability-cli-guides
Query metrics, traces, logs, and profiles from observability backends via their CLIs.
using-systemget-status
Render the ODD loop status from committed .odd/ reports and git metadata.
using-systemsetup-local-stack
Configure gcx against the local oddyssey Grafana stack and inventory its four telemetry signals.
using-systemrun-scenario
Drive reproducible HTTP request scenarios against local services and record them verbatim for telemetry comparison.
using-systempackage-layout
Reports the installed package's skills root, sibling directories, and each skill's files.
using-systembackend-configuration
Verify and switch the configured observability backend CLI before dispatching missions.
using-systemtest-plugin-harnessing
Measure and optimize one phase of an oddyssey run under a coding-agent CLI.
Frequently Asked Questions About Matthieu Vlad
FAQPage SchemaWhat tasks can I accomplish with using-system's skills?▼
You can plan OpenTelemetry instrumentation per language, author and replay k6 load benchmarks, drive reproducible request scenarios against local services, query metrics/traces/logs/profiles from Grafana, Datadog, Dynatrace, Azure, or AWS CLIs, and render the ODD loop state offline from committed .odd/ history and git.
Who are these skills designed for?▼
DevOps engineers, cloud architects, and SREs who instrument services with OpenTelemetry, run k6 performance benchmarks, and operate observability backends. It also targets maintainers who must prove harnessing changes on /odd-observe, /odd-verify, or /odd-status paths with measured evidence before opening a PR.
How do I set up the local observability stack?▼
The setup-local-stack skill configures gcx against the local oddyssey Grafana stack (default port 3000) and resolves Tempo, Prometheus, Loki, and Pyroscope datasource UIDs. gcx is mandatory; install it via brew install gcx or the official script from github.com/grafana/gcx.
What are the prerequisites and dependencies?▼
You need the stack-specific query CLI for your backend: gcx for Grafana, Pup CLI for Datadog, dtctl for Dynatrace, az for Azure Monitor, or aws for CloudWatch/X-Ray, plus k6 for benchmarks. The backend-configuration skill verifies installation, offers guided installs, and never installs silently or stores secrets.
How does the package measure and optimize its own run performance?▼
The test-plugin-harnessing skill drives coding-agent CLIs (opencode, claude, copilot) headless, measures one run phase (preflight, drive, observation, or full run) against main, names where time went, and separates package-controlled cost from provider latency. It grades the harness, not findings.