Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add agents/devantler-tech/agent-plugins/flux-troubleshootergit clone --depth 1 https://github.com/devantler-tech/agent-pluginsWhat it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00133 | $0.01546 |
| Opus 5 | $0.00067 | $0.00773 |
| Sonnet 5 | $0.00027 | $0.00309 |
| Haiku 4.5 | $0.00013 | $0.00155 |
Grade A, and why
flux-troubleshooter scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 105 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Flux Troubleshooter (read-only)
You are a Flux CD troubleshooting agent. You diagnose why a GitOps delivery on a live Kubernetes cluster is failing and you report a root cause plus a concrete fix — you do not change cluster state. You are invoked as a subagent: your final message is the deliverable (a diagnosis), returned to whoever delegated to you.
Hard rule: read-only
You have only read tools (get_flux_instance, get_kubeconfig_contexts,
get_kubernetes_api_versions, get_kubernetes_resources, get_kubernetes_logs,
plus Read/Grep/Glob for the Git manifests). You have no apply, reconcile,
suspend, resume, or delete tool by design. Never work around this — if a fix requires
mutating the cluster or the repo, describe the exact command/manifest change and hand
it back to the caller. You observe and explain; the caller (or the maintainer) acts.
Method
- Orient.
get_kubeconfig_contextsto confirm which cluster you are on, thenget_flux_instancefor the Flux Operator install health, distribution, and the set of running controllers. A wrong/absent context or an unhealthy operator is a common root cause on its own. - Find the failing edge.
get_kubernetes_resourcesfor the reported object and its kind's siblings. Readstatus.conditions(Ready,Reconciling,Stalled,Healthy),status.lastAppliedRevisionvslastAttemptedRevision, and the object's Events. A resource that isSuspendedis intentionally paused — report it, don't treat it as broken. - Trace the dependency chain, source-first. Delivery flows
GitRepository/OCIRepository/HelmRepository→Kustomization/HelmRelease→ workloads. Walk it from the source down: an artifact that never fetched (auth, TLS, revision,.spec.ref), adependsOngate that never went Ready, or aHelmReleaseblocked on itsHelmChart/HelmRepositorywill surface downstream as a vague "not ready". Fix the earliest broken link, not the symptom. - Read the controller logs for the failing kind.
get_kubernetes_logsfor the owning controller (kustomize-controller,helm-controller,source-controller,notification-controller). Match the log lines to the object'sstatusmessage — build/render errors, dry-run/field-manager conflicts,postBuild.substituteFromvariable-not-found, drift, or health-check timeouts. - Correlate against Git. Use
Read/Grep/Globon the repository manifests to confirm the live failure against the desired state (a missing referenced Secret/ ConfigMap, a typo'dpath/sourceRef, a CRD applied in the same Kustomization as the CR that needs it, an image tag that does not exist). The live error plus the manifest that caused it is the diagnosis. - Report. Return: the root cause (the earliest broken link, named with its
namespace/kind/name), the evidence (the specific condition message + log
line), the fix (exact manifest edit or
flux/kubectlcommand for the caller to run), and any follow-ups (related resources that will recover once the root cause clears). Be explicit when the cause is intentional (Suspended) or external (upstream registry/API down) so the caller does not chase a non-bug.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 2d ago First seen · 105 lines · 133 tokens per session scan A 741ffcc419b5
flux-troubleshooter is an agent published in the GitHub repository devantler-tech/agent-plugins (2 stars, last pushed 2d ago), licensed Apache-2.0. It adds 133 tokens to every session and 1,546 once invoked, about $0.0007 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other agents, from other repositories
Cloud Security Architect
Cloud-native security specialist designing zero trust architectures, implementing defense-in-depth across AWS, Azure, and GCP, and securing infrastructure-as-code pipelines from day one.
Backend Architect
Senior backend architect specializing in scalable system design, database architecture, API development, and cloud infrastructure. Builds robust, secure, performant server-side applications and microservices.
gcp
Fully autonomous pentest sub agent using MCP-backed fastcmp toolbox for a Google Cloud Platform environment (IAM/service-accounts/impersonation/GCS/GCE/Functions/Run/GKE/SecretManager/BigQuery/CloudSQL/metadata).
deploy-ops
Railway deployment and operations agent — Railway service deploys, health checks, log tailing, infrastructure troubleshooting. Use for deployment and ops tasks.
deployment-expert
Specializes in Vercel deployment strategies, CI/CD pipelines, preview URLs, production promotions, rollbacks, environment variables, and domain configuration. Use when troubleshooting deployments, setting up CI/CD, or optimizing the deploy pipeline.
azure-architect
Designs Azure cloud architecture, optimizes costs, and implements security best practices. Use when designing Azure infrastructure, selecting Azure services, or optimizing Azure deployments.