log-capability-gap

Autonomous iteration loop for optimizing any measurable target.

the reference deployment standard for writing and editing canonical skills (SKILL.md) in the Brain.

Evaluate whether a the reference deployment skill actually works — blind runner + separate judge against a bar, rubric scoring, pass-rate on re-runs, and…

Turn a big or ambiguous technical task into a spec of independently verifiable slices, implement slice-by-slice with verification gates, and close the spec…

Verify visual work with evidence instead of vibes — screenshot critique by a fresh judge, before/after comparison, crops of the disputed detail, and a…

Transfer full working context to a fresh agent session with zero memory — end of session, before compaction, switching models/agents mid-project, or pausing…

Use when an output produced by a human/agent needs to be checked, linked back to its task, and moved through review/done.

Never re-fight a settled battle.

Write research prompts that come back decision-ready — zero-prior-knowledge framing, evidence hierarchy, fact/inference separation, and a mandatory gap round.

Use when captured Brain context needs to become executable work.

Use when a Brain task card should be delegated to Claude, Codex, or a swarm agent with all context included.

Detecta SKUs the company con alta velocidad de venta + low stock y genera propuesta automatizada de restock con forecast para the merchandising agent + the…