Desenvolvimento

Extract uncaptured knowledge from the current conversation before it's lost to compaction. Use after completing a task, before switching context, before large exploratory work (multi-file reads, codebase scans), or when the user signals session end. Trigger: '/extract', 'extract knowledge', 'capture session knowledge'. Also prompt mid-session: 'Task complete — want me to run /extract?' and 'Switch

Structured retrospective on a shipped commit range, release, PR, commit, or session: per-fix validation, evidence-sourcing pass, re-diagnosis, action verdicts, growing failure-mode library. Triggers — "/aria-cowork:retrospect", "retrospect this". (Cowork variant — namespaced-only.)

Run a 4-step validation probe for the aria-cowork spec. Tests whether this Cowork plugin can verify cwd, write to the user-attached knowledge folder, read a file pre-placed by aria-knowledge in Code, and capture or fall back gracefully on the transcript surface. Outputs structured results both to the conversation and to a probe-results file. Use when validating the aria-cowork spec before Phase 1

Show knowledge base health metrics — file counts, backlog depth, audit status, tag stats, and coverage gaps. Use when user says "/aria-cowork:stats", "knowledge stats", "how is my knowledge base", "show stats", "knowledge health", "dashboard". (Cowork variant — namespaced-only.)

Mirror approved ARIA decisions from the knowledge folder to connected docs MCP destinations with explicit external-write review. Trigger on /sync-decisions or sync ADRs to docs.

Run a structured pre-mortem on a plan or approach BEFORE execution. Per-step risk enforcement, active evidence-sourcing pass (autonomous lookups + targeted user-asks for anything that could become objective), simpler-alternative discipline, plan-formation diagnosis, action verdicts (PROCEED/SHRINK/SPLIT/DEFER/KILL), and a growing failure-mode pattern library. Triggers: '/prospect' (defaults to pla

First-message audit substitute for Cowork. Use when user says "/daily-audit", "/aria-cowork:daily-audit", "audit check", "daily check", "session start audit", or at the start of a Cowork session to surface audit-cadence status. Cowork has no SessionStart hook (per ADR-004); this skill is the manual equivalent — checks /audit-knowledge +.

Read subtask.md checkbox state and route to next skill. The ONLY routing point in the execution loop. Called after autoworker:checkpoint, autoworker:gate-check, autoworker:subtask-update, autoworker:subtask-plan, or when resuming after context loss. When lost, call autoworker:dispatch.

Use when creating, modifying, auditing, or debugging Metin2 client UI code (uiscript dicts, root ui*.py classes, locale entries) — user says "/m2ui", "create UI", "make a window", "add button to", provides a UI screenshot to replicate, asks to "check UI for bugs" / "diagnose", or reports a visibly broken window ("doesn't open", "click does nothing", "leaks").

Show ARIA knowledge-base health metrics: counts, backlog depth, audit status, codemap freshness, tag coverage, and gaps. Trigger on /stats or knowledge health.

Manually clean up development track worktrees. Debug/Expert command - rarely needed as worktrees are cleaned up automatically.

Build a cross-tool digest from connected chat, email, project tracker, and docs MCPs, with graceful fallback when connectors are missing. Trigger on /digest or weekly digest.