rocky2431

ai-collab-base

Shared collaboration protocol, modes, and prompt templates for codex-collab and ultra-verify.

learn

Extract one reusable pattern from the current session into a valid user skill, with explicit user approval before writing.

Create or recover Ultra's repository-local skeleton and raw Project Brief without performing product research. Use when `.ultra` is absent or partial in a new or existing codebase and later workflows need stable file paths, preserved authority, and Git.

Code & Development Listed

code-review-expert

Review a bounded diff for consequential correctness, security, integration, and maintainability defects with evidence and precise locations. Use only when assigned to an Ultra review worker or an explicitly requested read-only code review.

ultra-deliver

Reconcile one Change whose matching tasks are complete, run aggregate review and refactoring, and report the delivery handoff. Use when the owner wants delivery reconciliation and reporting, or — by current explicit owner invocation only — the separately gated finalization that writes delivery.md, fixes version and package posture, and archives the local workflow record.

ultra-dev

Carry one planned task from a written implementation plan through red-green development to recorded evidence and a task-level review. Use when a task in the ledger is ready to execute, was interrupted partway, or has to be picked up in a fresh session or on another host.

ultra-plan

Turn an evidenced Change intent into a dependency-valid task ledger, confirmed test seams, and resumable task contexts. Use when implementation needs decomposition, a wide refactor needs an expand-contract route, or an interrupted plan must be reconstructed from repository files.

ultra-research

Turn a raw Project Brief or scoped evidence gap into an accepted North Star, domain language, cited product and architecture specifications, and a hash-bound synthesis. Use after initialization, for brownfield baselining, or inside an active Change whose consequential claims are not yet evidenced.

ultra-verify

Run independent two-model verification with the current host as primary and one installed read-only advisor, then verify and synthesize both evidence sets. Use only when the user explicitly requests cross-model verification.

ultra-change

Open or reconcile one bounded unit of work by comparing specification promises with actual code, then writing the intent that closes the gap. Use when no Change is active and a feature, fix, redesign, or incident response is requested, or when an active Change's draft, researched evidence, or accepted boundary must be reconciled.

Code & Development Listed

ultra-review

Review one task diff or aggregate Change through owner-selected independent lenses, synthesize file-backed findings, and perform evidence-justified refactoring. Use when another skill needs task-level or delivery-level review without loading worker investigation into the main context.

ultra-status

Infer the current Ultra route, baseline gaps, stale evidence, task risk, and installation health from repository files and host manifests. Use when the owner asks what is complete, blocked, stale, unhealthy, or useful to do next.

Testing & QA Listed

ultra-test

Audit whole-system integrity after tasks for the active Change are complete by finding anti-patterns, coverage gaps, broken wiring, E2E failures, performance risk, and security risk. Use when preparing delivery or when locally green tasks may not form a working product.

ultra-think

Stress-test one consequential product, architecture, recovery, or scope decision and preserve the accepted result when it must outlive the session. Use when another skill has one evidence-backed trade-off whose answer could materially change accepted work.

ultra-delegate

Delegate one bounded workflow segment to another installed CLI in an isolated worktree and return only its stable result artifact. Use when the owner asks for another model or host to execute, review, verify, or continue a specific slice of current work.

ultra-domain-modeling

Settle the project's vocabulary into CONTEXT.md — one opinionated definition per term, the rejected wordings, the relationships between terms, and the ambiguities already resolved. Use when another skill changes the shared language, meaning a term gets coined, two words compete for one concept, or one word turns out to mean two things.

ultra-grilling

Turn caller-named missing owner decisions into confirmed answers, one question per turn and each carrying a recommendation. Use when Init needs a raw brief field, Research needs an owner decision, or Change and Plan need an explicit boundary or acceptance criterion.

Testing & QA Listed

ultra-tdd

Drive one slice of behavior through red then green — a failing test on a confirmed seam, then the smallest code that passes it. Use when another skill is about to write or change behavior and needs the feedback loop that proves the code actually runs, including a small change that goes straight to implementation.

integration-rules

Verify that changed components are reachable, contract-compatible, and tested across the boundaries required by the accepted behavior. Use only when an assigned review or implementation slice crosses modules, services, persistence, protocols, or UI seams.

security-rules

Evaluate changed trust boundaries for exploitable input, authorization, secret, data-exposure, and supply-chain failures. Use only when an assigned review scope contains a security-relevant boundary.

cc-collab

Ask Claude Code for an independent read-only analysis of a bounded architecture, diagnosis, or review question. Use only when the user explicitly requests Claude Code or CC collaboration and the current host remains primary.

Testing & QA Listed

testing-rules

Review whether changed behavior has meaningful, trustworthy, and proportionate test evidence across its real boundaries. Use only when assigned to a test review or delivery verification scope.

codex-collab

Ask Codex CLI for an independent read-only analysis of a bounded architecture, diagnosis, or review question. Use only when the user explicitly requests Codex collaboration and the current host remains primary.

ultra-doctor

Diagnose and mechanically recover Ultra schema, projection, session, archive, hook, and installed-asset faults. Use when authority is degraded, contradictory, or unavailable.