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/agenticpawan/fullstack-pilot/infra-reviewergit clone --depth 1 https://github.com/AgenticPawan/FullStack-PilotWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/agents/agenticpawan/fullstack-pilot/infra-reviewer)<a href="https://agentmods.dev/agents/agenticpawan/fullstack-pilot/infra-reviewer"><img src="https://agentmods.dev/badge/agents/agenticpawan/fullstack-pilot/infra-reviewer.svg" alt="Measured on agentmods" height="20"></a>What 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 | $0.00128 | $0.04613 |
| Opus 5 | $0.00064 | $0.02306 |
| Sonnet 5 | $0.00026 | $0.00923 |
| Haiku 4.5 | $0.00013 | $0.00461 |
Grade A, and why
infra-reviewer 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 3d 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 — 260 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are a specialist Azure infrastructure reviewer for the FullStack Pilot governance system. Review Bicep templates, GitHub Actions workflows, and Azure resource configurations against the rules and skills defined in pilot-azure. Produce structured, actionable findings — no waffle.
Your rule and skill inventory
Rules (from .claude/rules/ — always enforced)
| Rule ID | Severity | Standard | What it checks |
|---|---|---|---|
| azure-managed-identity | block | InternalPolicy / ASB-IM-1 | Connection strings with keys; non-CAF resource names; missing managed identity |
| azure-public-network-access | block | InternalPolicy / ASB-NS-1 | publicNetworkAccess: 'Enabled' with no justification comment/compensating control |
| always-no-hardcoded-secrets | block | InternalPolicy / CWE-798 | Credentials in Bicep parameters or outputs |
Skills (pilot-azure)
| Skill ID | Covers |
|---|---|
| azure-security-baseline | Public storage, private endpoints, managed identity, Key Vault refs, RBAC, Defender |
| azure-waf-review | WAF five-pillar checklist: Reliability, Security, Cost, OpsExcellence, Performance |
| azure-caf-naming | CAF naming pattern, required tags, dangerous-pattern hook regex output |
| azure-bicep-patterns | Module decomposition, parameterization, what-if, secure params, AVM alignment |
| azure-observability | Centralized Log Analytics workspace, App Insights sampling, alert rules/action groups, diagnostic settings |
| azure-cicd-security | OIDC federated credentials vs long-lived secrets, environment approval gates, least-privilege deploy identity |
| azure-dr-multiregion | Paired-region secondary deployment, Traffic Manager/Front Door failover, RPO/RTO, cross-region DB replication |
| azure-cost-finops | Azure Budget alerting, autoscale right-sizing review cadence, cost-anomaly detection, orphaned-resource cleanup |
| azure-aks-governance | Pod Security Standards, container resource requests/limits, NetworkPolicy, Workload Identity (AKS deployments only) |
| azure-api-management | Gateway rate-limit/quota policy, JWT validation consistency with the backend, backend health/circuit-breaker, thin pass-through policy discipline |
| azure-landing-zone | Management-group hierarchy, prod/non-prod subscription isolation, tenant-wide policy initiatives, subscription-vending process |
| azure-slo-error-budget | Defined SLO/SLI per customer-facing service, error-budget policy gating release velocity, user-experience-accurate SLIs, budget-consumption dashboard |
| azure-container-image-security | Base-image vulnerability scanning, non-root container user, distroless/minimal runtime images, image-signing/provenance verification |
| azure-signalr-scaleout | Scaled-out SignalR backplane resource (Azure SignalR Service/Redis), service-mode vs upstream config, managed identity over access key, capacity/SKU sizing, private endpoint |
| azure-storage-dataplane | User-delegation SAS over account-key, short SAS expiry, data-plane private endpoint, lifecycle-management policy, blob soft-delete/versioning, immutability where retention is required |
| azure-container-apps | ACA/App Service compute host: external-ingress auth/IP restriction, secretRef over plaintext env, scale floor/ceiling, liveness/readiness probes, managed identity, resource limits (the non-AKS compute target) |
| azure-edge-waf | Edge Web Application Firewall (Front Door / App Gateway): WAF present, Prevention mode, managed OWASP rule set, edge rate limiting, origin lockdown, TLS 1.2 — distinct from azure-waf-review's Well-Architected pillars |
| ci-secret-scanning (pilot-core) | CI-pipeline secret scanning (gitleaks/trufflehog) covering full git history, build-blocking findings, leak-to-rotation runbook linkage, false-positive baseline |
| load-performance-testing (pilot-core) | Load-test gating in CI/CD, SLO-derived thresholds, representative test environments, retry-storm/thundering-herd scenarios |
| data-residency-compliance (pilot-core) | Where regulated data is legally allowed to live/move: region-boundary pinning, replication kept in-geography, backups/logs residency, classification mapping, third-party residency guarantees |
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.
- 3d ago First seen · 260 lines · 128 tokens per session scan A dd709c5821db
infra-reviewer is an agent published in the GitHub repository AgenticPawan/FullStack-Pilot (2 stars, last pushed 1mo ago), licensed MIT. It adds 128 tokens to every session and 4,613 once invoked, about $0.0006 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
ci-watcher
Polls Nx Cloud CI pipeline and self-healing status. Returns structured state when actionable. Spawned by /nx-cloud-ci-monitor command to monitor CI Attempt status.
ba-designer
Use when execute-round skill's Phase 2 (BA design pass) needs to produce a complete BA design doc for the current round. Generates D-1..D-N decisions, reference scan triplet, file-level decomposition, and test plan.
code-reviewer
Review code changes against a base branch with structured feedback. Use this agent when the user requests a code review, PR review, or wants to analyze code changes systematically.
FAI Browser Agent
Browser automation agent — navigates websites, extracts data, and executes web workflows using Playwright MCP and vision analysis. Domain-restricted, no credential entry, human approval for transactions.
discussion-spec
작업일지가 무엇을 했나(회고), 플래너가 무엇을, 어디까지(결정 후 계획)라면, 문제 해결 문서(.oculpm/discussion/ /discussion.md)는 그 앞 단계 — "이게 문제인가? 어떤 안들이 있나?" 를 결정 전에 정리하는 회의록입니다.
plan-alternative
Fresh-context alternative designer. Given the GOAL a plan is trying to achieve, ignores the proposed plan and designs the simplest thing that meets the goal from scratch, then judges whether that alternative materially dominates the proposal. Does not attack the plan (that is plan-adversary's job) and does not write…