{
  "markdown": "# Staff Engineer Mode\n\n**Staff-level engineering judgment for coding agents, from design to production.**\n\nGive Staff Engineer Mode a design, diff, rollout, incident, migration, or\nmaintenance problem. One router selects one of 64 focused specialists. The\nagent works through relevant failure modes and verification checks before it\nships code or changes production.\n\nYou get concrete decisions, risks, checks, and next steps across\narchitecture, reliability, security, delivery, data, platform, and operations.\n\n## How It Works\n\nAsk a normal engineering question. You never need to name a specialist.\n\nThe router classifies the requested artifact, lifecycle phase, engineering\nsurface, and risk. It chooses one primary specialist and adds a secondary only\nwhen the request needs a separate artifact.\n\nSupported tools list only the native `staff-engineer-mode` router. The 64\nspecialist files stay under `specialists/` and load only after routing.\n\nFor commits and amends, Staff Engineer Mode calls `agent-pr-review` against\nthe exact staged diff. For releases, tags, version bumps, packages, artifacts,\nand promotions, it calls `release-build-reproducibility` and\n`production-readiness-review` together.\n\n## Installation\n\nCommands labeled \"terminal\" are run in your shell. Commands labeled \"agent\nchat\" are typed inside that tool's interactive agent session.\n\n### Claude Code\n\nTerminal:\n\n```bash\nclaude plugin marketplace add https://github.com/sirmarkz/staff-engineer-mode.git\nclaude plugin install staff-engineer-mode@staff-engineer-mode\n```\n\nAgent chat:\n\n```text\n/plugin marketplace add https://github.com/sirmarkz/staff-engineer-mode.git\n```\n\n```text\n/plugin install staff-engineer-mode@staff-engineer-mode\n```\n\n### Codex\n\nTerminal:\n\n```bash\ncodex plugin marketplace add https://github.com/sirmarkz/staff-engineer-mode.git\ncodex plugin add staff-engineer-mode@staff-engineer-mode\n```\n\n### Cursor\n\nTerminal:\n\n```bash\ngit clone https://github.com/sirmarkz/staff-engineer-mode.git ~/.cursor/staff-engineer-mode-src\nmkdir -p ~/.cursor/plugins\nln -s ~/.cursor/staff-engineer-mode-src ~/.cursor/plugins/staff-engineer-mode\n```\n\n### OpenCode\n\nTerminal:\n\n```bash\nopencode plugin 'staff-engineer-mode@git+https://github.com/sirmarkz/staff-engineer-mode.git'\n```\n\n### GitHub Copilot CLI\n\nTerminal:\n\n```bash\ncopilot plugin marketplace add https://github.com/sirmarkz/staff-engineer-mode.git\n```\n\nInstall the plugin:\n\n```bash\ncopilot plugin install staff-engineer-mode@staff-engineer-mode\n```\n\n### Gemini CLI\n\nTerminal:\n\n```bash\ngemini extensions install https://github.com/sirmarkz/staff-engineer-mode\n```\n\n## Try It\n\nStart a fresh session inside any open repo and try one of these prompts:\n\n- \"Before implementing partner webhooks, design delivery retries, replay, and dead-letter handling.\"\n- \"For a new inventory dependency call, decide timeout, retry, and fallback.\"\n- \"Review my last commit.\"\n\nThe agent should load the router, choose one specialist, and respond with\nconcrete decisions, risks, checks, owners, supporting details, and next steps.\n\nFor more coverage, see the [sample prompts](evals/prompts/expected-routes.md).\n\n## What's Inside\n\nStaff Engineer Mode ships one native router skill. It keeps 64 specialist files\nunder `specialists/` and loads them after routing.\n\nThe concern groups below help you browse. Runtime routing still follows the\nrequested artifact, phase, surface, and risk.\n\n| Engineering concern | Specialist files |\n| --- | --- |\n| Architecture & interfaces | [`architecture-decisions`](specialists/architecture-decisions.md), [`api-design-and-compatibility`](specialists/api-design-and-compatibility.md), [`data-contracts`](specialists/data-contracts.md), [`event-workflows`](specialists/event-workflows.md), [`resilience-requirements`](specialists/resilience-requirements.md), [`persistent-connection-systems`](specialists/persistent-connection-systems.md) |\n| Verification & evaluation | [`state-machine-correctness`](specialists/state-machine-correctness.md), [`testing-and-quality-gates`](specialists/testing-and-quality-gates.md), [`test-data-engineering`](specialists/test-data-engineering.md), [`agent-pr-review`](specialists/agent-pr-review.md), [`experimentation-and-metric-guardrails`](specialists/experimentation-and-metric-guardrails.md) |\n| Reliability & resilience | [`slo-and-error-budgets`](specialists/slo-and-error-budgets.md), [`high-availability-design`](specialists/high-availability-design.md), [`dependency-resilience`](specialists/dependency-resilience.md), [`backup-and-recovery`](specialists/backup-and-recovery.md), [`resilience-experiments`](specialists/resilience-experiments.md), [`performance-and-capacity`](specialists/performance-and-capacity.md), [`cost-aware-reliability`](specialists/cost-aware-reliability.md), [`multi-region-and-data-residency`](specialists/multi-region-and-data-residency.md), [`scheduled-job-reliability`](specialists/scheduled-job-reliability.md) |\n| Data, storage & privacy | [`distributed-data-and-consistency`](specialists/distributed-data-and-consistency.md), [`database-operations`](specialists/database-operations.md), [`data-pipeline-reliability`](specialists/data-pipeline-reliability.md), [`caching-and-derived-data`](specialists/caching-and-derived-data.md), [`privacy-and-data-lifecycle`](specialists/privacy-and-data-lifecycle.md), [`data-lineage-and-provenance`](specialists/data-lineage-and-provenance.md) |\n| Delivery & change safety | [`progressive-delivery`](specialists/progressive-delivery.md), [`feature-flag-lifecycle`](specialists/feature-flag-lifecycle.md), [`release-build-reproducibility`](specialists/release-build-reproducibility.md), [`fleet-upgrades`](specialists/fleet-upgrades.md), [`migration-and-deprecation`](specialists/migration-and-deprecation.md), [`configuration-and-automation-safety`](specialists/configuration-and-automation-safety.md), [`dev-environment-parity`](specialists/dev-environment-parity.md), [`service-decommission-and-sunset`](specialists/service-decommission-and-sunset.md) |\n| Code quality & maintainability | [`code-readability-for-agents`](specialists/code-readability-for-agents.md), [`dependency-and-code-hygiene`](specialists/dependency-and-code-hygiene.md) |\n| Operations & incident response | [`observability-and-alerting`](specialists/observability-and-alerting.md), [`incident-response-and-postmortems`](specialists/incident-response-and-postmortems.md), [`oncall-health`](specialists/oncall-health.md), [`operational-ownership-transfer`](specialists/operational-ownership-transfer.md) |\n| Security | [`secure-sdlc-and-threat-modeling`](specialists/secure-sdlc-and-threat-modeling.md), [`identity-and-secrets`](specialists/identity-and-secrets.md), [`cryptography-and-key-lifecycle`](specialists/cryptography-and-key-lifecycle.md), [`software-supply-chain-security`](specialists/software-supply-chain-security.md), [`vulnerability-management`](specialists/vulnerability-management.md), [`tenant-isolation`](specialists/tenant-isolation.md), [`edge-traffic-and-ddos-defense`](specialists/edge-traffic-and-ddos-defense.md), [`llm-application-security`](specialists/llm-application-security.md), [`input-validation-and-injection-defense`](specialists/input-validation-and-injection-defense.md), [`client-application-security`](specialists/client-application-security.md) |\n| Platform & infrastructure | [`infrastructure-and-policy-as-code`](specialists/infrastructure-and-policy-as-code.md), [`internal-service-networking`](specialists/internal-service-networking.md), [`platform-golden-paths`](specialists/platform-golden-paths.md), [`container-runtime-and-orchestration`](specialists/container-runtime-and-orchestration.md) |\n| Client applications & accessibility | [`web-release-gates`](specialists/web-release-gates.md), [`mobile-release-engineering`](specialists/mobile-release-engineering.md), [`accessibility-gates`](specialists/accessibility-gates.md) |\n| AI/ML evaluation & serving | [`llm-evaluation`](specialists/llm-evaluation.md), [`ml-reliability-and-evaluation`](specialists/ml-reliability-and-evaluation.md), [`llm-serving-cost-and-latency`](specialists/llm-serving-cost-and-latency.md) |\n| Engineering controls & readiness | [`ai-coding-governance`](specialists/ai-coding-governance.md), [`documentation-lifecycle`](specialists/documentation-lifecycle.md), [`engineering-control-evidence`](specialists/engineering-control-evidence.md), [`production-readiness-review`](specialists/production-readiness-review.md) |\n\n## Sources\n\nThe practice library uses first-party engineering publications from Amazon,\nGoogle, Meta, Microsoft, Apple, and Netflix. Standards and guidance come from\nNIST, CISA, OWASP, OpenSSF, IETF, and W3C.\n\nThe library covers public incident records from AWS, Azure, Google Cloud,\nGoogle Workspace, Meta, and Netflix.\n\nSee the [source index](skills/_shared/references/source-index.md) for the full\nreference set. Staff Engineer Mode is independent and is not endorsed by or\naffiliated with these organizations.\n\n## Contributing\n\nPatches are welcome. New practices should come from authoritative sources:\nfirst-party engineering publications, official documentation, standards\nbodies, peer-reviewed papers, or widely cited practitioner references.\n\nContributors must keep new specialists technology-agnostic. Do not endorse\nvendors. Maintain authoritative sources in the shared index and read\n[CONTRIBUTING.md](CONTRIBUTING.md) before opening a PR. Repository checks\nenforce this voice.\n\n## License\n\nMIT\n",
  "bytes": 9487,
  "sha": "82834136b7af73bbb8f8ca505be777c9afe7e624e3e538e36ef0c6c2d0929003",
  "repo_slug": "sirmarkz/staff-engineer-mode",
  "fonte": "repo",
  "truncated": false,
  "api": "https://agentalog.com/api/listings/plg_sirmarkz_staff_engineer_mode_staff_engin_12322edb/readme"
}