Your AI engineering org, governed. Ship agent-built code safely to production.
Agent Claude Code
Implements domain logic, APIs, and infrastructure following clean architecture principles. Activates during the Build phase on backend code (domain / application / infrastructure layers), API work, or database schema changes.
Agent Claude Code
Expert code review specialist. Reviews PRs for quality, security, and standards compliance. Use proactively after code changes or when a PR needs review.
Agent Claude Code
The Contrarian — advisory adversarial reviewer of IDEAS, not artifacts. Steelmans then challenges a feature, spec, decision, plan, or AgDR — surfacing hidden assumptions, failure modes, cheaper alternatives, and a proceed / proceed-with-changes / reconsider verdict. Invoked on demand via /challenge or "play devil's…
Agent Claude Code
Writes SQL, builds dashboards, runs A/B-test analysis, and investigates metrics. Activates on SQL / dashboard / A/B-test / metric-investigation work — quantitative, fast, narrow tool-use (candidate for local-model routing per.
Agent Claude Code
Builds ETL pipelines, designs data models, owns data-quality work, and manages warehouse schema changes. Activates on ETL / data-modelling / warehouse-schema / data-quality work — pipeline implementation, often in-flow with Backend Engineer handoff.
Agent Claude Code
Monitors dependencies for vulnerabilities, outdated packages, and license compliance. Run weekly or when package.json changes.
Agent Claude Code
Builds user interfaces following the design system — accessible, performant, and delightful. Activates during the Build phase on UI code, component work, design-system integration, or accessibility review.
Agent Claude Code
Owns analytics strategy, data governance, reporting architecture, and cross-project data modelling. Activates on cross-project data calls, governance decisions, reporting-architecture reviews, and strategic data tooling choices.
Agent Claude Code
Owns the design system, UX principles, and visual standards across products. Activates on design-system changes, UX principles decisions, or cross-project visual-standards calls.
Agent Claude Code
Owns engineering strategy, architecture standards, and engineering culture across the portfolio. Activates on architecture review, new tech stack additions, cross-project engineering calls, and Tech Lead escalations.
Agent Claude Code
Owns product strategy, roadmap prioritisation, and feasibility studies. Activates on roadmap prioritisation, feasibility calls, strategic product decisions, or cross-product resource allocation.
Agent Claude Code
Owns security strategy, threat-model gating, compliance calls, and cross-project security architecture. Activates on strategic security decisions, organisation-level threat-model reviews, compliance-deadline calls, and escalations from the Security Auditor.
Agent Claude Code
Adversarial security testing, exploit discovery, API security review, and pre-release security sign-off. Activates on explicit invocation for active testing, exploit reasoning, or hardening checks; not auto-fired on diff content.
Agent Claude Code
Builds and maintains infrastructure, CI/CD pipelines, and developer tooling — the platform that lets engineers ship fast and safely. Activates on CI/CD pipeline changes, developer tooling, infrastructure-as-code work, or golden-path templates.
Agent Claude Code
Provides data-driven insights, market research, and competitive analysis to support product decisions and feasibility studies. Activates on market research, competitive analysis, metric investigation, or data-driven product calls.
Agent Claude Code
Translates approved product strategy into detailed PRDs with acceptance criteria, coordinates with Design and Engineering, and removes delivery blockers. Activates on PRD creation, user-story breakdown, acceptance-criteria authoring, or sprint planning.
Agent Claude Code
Verifies acceptance criteria on merged PRs, triages bugs, runs regression checks, and signs off tickets before they move to Done. Activates when a ticket enters the QA state after merge. Read-only by design — QA verifies, doesn't ship.
Agent Claude Code
Security Auditor — runs OWASP / threat-model / SAST analysis on PR diffs and provides remediation guidance. Auto-activates on PRs touching auth, crypto, secrets, user data, APIs, third-party integrations, or the security-critical trust chain (.claude/hooks/, .claude/settings.json — the.
Agent Claude Code
Solution Architect — independent design reviewer. Reviews technical designs, migration AgDRs, and feature specs BEFORE the Build phase for architectural soundness (NFRs, patterns, tech debt, decisions, risk, trade-offs, traceability). The non-code analog of the Code Reviewer (Rex). Auto-activates on PRs that touch…
Agent Claude Code
Ensures systems are reliable, observable, and resilient — applies engineering to operational problems. Activates on production incidents, SLO breaches, monitoring / alerting work, or on-call rotation.
Agent Claude Code
Bridges architecture and implementation — authors technical designs, leads code reviews, mentors engineers, and owns technical quality for a domain. Activates on technical design, planning phase, code review approval gate, or task breakdown.
Agent Claude Code
Defines the visual language and component specifications that guide UI implementation. Activates on visual design, component specifications, design tokens, or pixel-level work.
Agent Claude Code
Focuses on user flows, information architecture, and usability — documenting journeys and ensuring products are intuitive and efficient. Activates on user flows, information architecture, usability review, or wireframing.