Dev skills

Skills are markdown playbooks the agent loads on demand (or automatically via workflow commands). Manage them in Settings → Skills or list them with /skill. The skills below power the Dev module; the full catalog contains many more (marketing, SEO, documents, HR, sales…).

Engineering

Skill	Purpose
`fullstack-dev`	End-to-end feature building: frontend, backend, database.
`api-engineer`	API design and implementation.
`task-planner`	Breaking work into ordered, verifiable steps (used by `/blueprint`).
`test-generator`	Writing meaningful test suites.
`code-reviewer`	Structured code review methodology (used by `/inspect`).
`critic-reviewer`	Adversarial second-pass review.
`skill-creator` / `skill-vetter`	Author and audit new skills.
`pack-builder`	Create and audit plugin packs.

Security

Skill	Purpose
`security-auditor`	Systematic security audits (used by `/fortify`).
`vulnerability-scanner`	Exploit hunting: OWASP, CVEs, secrets (used by `/probe`).
`prompt-injection-defender`	Detecting and defusing prompt-injection surfaces.
`secrets-manager`	Safe handling of credentials and tokens.
`permission-guard`	Enforcing permission boundaries.
`audit-logger`	Traceability of sensitive operations.

Performance & quality

Skill	Purpose
`performance-auditor`	Profiling and optimization (used by `/turbo`).
`metrics-analyst`	Project metrics collection and scoring (used by `/pulse`).
`kpi-reporter`	KPI dashboards and reporting.
`quality-gate`	Multi-gate pass/fail assessment.

Ops & infrastructure

Skill	Purpose
`docker-operator`	Containers: build, run, debug.
`kubernetes-operator`	K8s deployments and troubleshooting.
`terraform-agent`	Infrastructure as code.
`cicd-agent`	Pipelines and continuous delivery.
`observability-agent`	Logs, traces, metrics wiring.
`backup-rollback`	Safe state snapshots and restores.
`shell-sandbox`	Sandboxed shell execution practices.
`tmux`	Long-running terminal session management.

Data

Skill	Purpose
`database-explorer`	Schema discovery and querying.
`sql-analyst`	SQL analysis and optimization.
`supabase-operator`	Supabase projects: DB, auth, storage.
`stripe-operator`	Stripe integration and operations.
`data-migration-agent`	Safe schema/data migrations.
`data-quality-agent`	Data validation and cleaning.

Agent autonomy

Skill	Purpose
`multi-agent-orchestration`	Coordinating subagents on a task.
`adaptive-reasoning`	Choosing the right depth of reasoning.
`model-router`	Picking the right model per task (pairs with `/pilot`).
`context-compressor`	Keeping long sessions within context.
`self-healing-retry`	Recovering from failed steps.
`proactive-agent`	Anticipating needs during long goals.
`human-approval`	Pausing for human sign-off on sensitive steps.
`memory`	Long-term memory conventions (pairs with Dream).

Skills are loaded automatically by workflow commands (e.g. /fortify preloads security-auditor, permission-guard, secrets-manager) or on request: just ask the agent to "use the code-reviewer skill".