CATWILLgh

60 mods across 1 repository, 2 stars between them.

CATWILLgh/MAINFRAME

Skill Claude CodeCodex

Delegate one agreed, bounded implementation block to MAINFRAME's project-scoped Pi engineer after the result, boundaries, acceptance criteria, and allowed checks are clear. Use new for a new block and resume for corrections to the active block. Do not use for requirements discovery, architecture decisions, open-ended…

2 11d ago A 76 tokens copy · 88% MIT

CATWILLgh/MAINFRAME

Skill Claude CodeCodex

Implement, debug, review, or plan server-side Python in FastAPI, Django, Flask, and established Python services. Use for HTTP APIs, business logic, authentication, PostgreSQL, SQLAlchemy, migrations, workers, realtime behavior, caching, object storage, external integrations, generated documents, observability, and…

2 11d ago A 95 tokens original MIT

CATWILLgh/MAINFRAME

Skill Claude CodeCodex

Perform the final independent readiness check for a consequential task before a prepared decision is presented or an implemented result is accepted. Use only inside the owning advisor profile for preparation-readiness and final-state reviews; do not use for routine work, implementation, or adversarial decision review.

2 11d ago A 60 tokens original MIT

CATWILLgh/MAINFRAME

Skill Claude CodeCodex

Apply the private evidence, source-selection, cross-checking, numeric-verification, and stopping method of mainframeresearcher. Use only when preloaded into that custom agent; do not expose this specialist method to primary skill discovery.

2 11d ago A 53 tokens original MIT

mainframe-secrets

53

CATWILLgh/MAINFRAME

Skill Claude CodeCodex

Safely discover and consume credentials already registered on this machine without reading, echoing, logging, copying, or persisting their values. Use when a command, remote service, HTTP request, SSH connection, deployment, or external CLI needs a credential or a registered service description.

2 11d ago A 60 tokens original MIT

CATWILLgh/MAINFRAME

Skill Claude CodeCodex

Design or audit a cross-cutting testing strategy when test levels, suite cost, infrastructure boundaries, reliability, or regression coverage need a deliberate decision. Use for test-suite audits, slow or misleading suites, coverage spanning several components, or deciding which guarantees belong in fast local tests…

2 11d ago A 85 tokens original MIT

mainframe-ticket

55

CATWILLgh/MAINFRAME

Skill Claude CodeCodex

Record a concrete problem outside the active task's assigned result or agreed definition of done without expanding scope. Use when work reveals an incidental observation worth preserving or an explicitly assigned investigation confirms a problem with evidence. Do not use to replace unfinished in-scope work, scope…

2 11d ago A 65 tokens original MIT

CATWILLgh/MAINFRAME

Skill Claude CodeCodex

Search the current local repository or a named scope broadly for concrete plausible problems and record deduplicated ticket observations without fixing or fully confirming them. Use only when the user explicitly starts a ticket-discovery run in native Goal mode. Do not use for ordinary incidental findings, focused…

2 11d ago A 68 tokens original MIT

CATWILLgh/MAINFRAME

Skill Claude CodeCodex

Implement explicitly autonomous ready tickets one at a time, establish focused red evidence where useful, validate each complete fix, commit coherent recovery points, and queue results for independent verification. Use only when the user explicitly starts a ticket-implementation run in native Goal mode for the…

2 11d ago A 83 tokens original MIT

CATWILLgh/MAINFRAME

Skill Claude CodeCodex

Verify, scope, atomize, and consolidate open ticket observations, routing each result to ready technical work, a user decision, rejection, or an explicit evidence gap. Use only when the user explicitly starts a ticket-refinement run in native Goal mode for all eligible tickets or a named scope. Do not use for…

2 11d ago A 80 tokens original MIT

CATWILLgh/MAINFRAME

Skill Claude CodeCodex

Independently verify implemented tickets, archive proven fixes, and return failed work to the correct open queue without repairing it inline. Use only when the user explicitly starts a ticket-verification run in native Goal mode from a task that did not implement the queued work. Do not use for discovery, refinement…

2 11d ago A 73 tokens original MIT

CATWILLgh/MAINFRAME

Skill Claude CodeCodex

Implement, debug, review, or plan server-side TypeScript in Node.js services, NestJS, Express, Fastify, and established Next.js server layers. Use for HTTP APIs, business logic, PostgreSQL, Prisma, TypeORM, Drizzle, auth, migrations, workers, queues, realtime communication, storage, resilience, observability, and…

2 11d ago A 100 tokens original MIT