Coding & Development

Review a proposed customer code change against its intended outcome and operational risks. Use for a diff or PR review; report evidence and actionable findings.

Conduct a retrospective on a coding session.

Assess a new customer request against agreed scope, trade-offs and ownership. Use when an engagement expands or a custom feature needs a commitment decision.

>-

Prepare or execute an authorized controlled release with verified checks, recovery and an operating owner. Use when a customer increment is ready for deployment.

Run a bounded customer proof of concept to test a consequential uncertainty. Use for a spike or pilot with a question and decision deadline, not a full rollout.

A runnable reproduction of EvoSkill automated skill discovery: learn reusable SKILL.md files from execution failures under a bounded top-K frontier. Measured on FinQA.

>-

>-

Delegate bounded source collection and return locally ranked inspectable evidence, or a worker handle for composition.

>-

Simplify structure, interfaces or responsibilities while preserving behavior. Use when: a focused refactor is requested; feature changes need their own intent.

Inspect disk pressure with SBH and perform an authorized recovery action. Use when: storage diagnosis or SBH recovery is requested; inspection does not authorize deletion.

Manage GitHub issues locally as Markdown files. Use for triaging, searching, editing, and creating issues without leaving your editor or terminal.

Review selected native agent history to improve the local environment, custom workflows, or Orchflows itself.

>-

Delegate one independent review of supplied evidence into a globally ranked, cited assessment.

Trigger: update skills, skill registry, actualizar skills, after skill changes. Index available skills by trigger and path.

Trigger: judgment day, judgement day, dual review, adversarial review, juzgar. Run explicit blind dual review with at most two scoped fix/re-judgment rounds.

Plan commits as reviewable work units. Trigger: implementation, commit splitting, chained PRs, or keeping tests and docs with code.

Trigger: /skill-creation, skill creation, skill creator, create skill, new skill. Create LLM-first skills with valid frontmatter.

Operate Gas City through its Mayor, registry packs and native run state. Use when: the caller explicitly selects Gas City; factory completion does not replace independent judgment.

Trigger: improve skills, audit skills, refactor skills, skill quality. Audit and upgrade existing LLM-first skills.

Review code or scan for security vulnerabilities, secrets, dependencies and prompt risks. Use when: concrete exposure needs assessment; never silently change policy.