coreycoto/product-development
Owned workflows for understanding, designing, implementing, testing, reviewing and verifying products.
Assess a specific dataset or derived artifact for its declared use using provenance, independently justified expectations, semantic checks, and explicit population or sampling limits without repairing source data.
Preserve verified non-obvious project reasoning whose loss could cause recurrence or substantial rediscovery, using the consumer’s existing knowledge convention.
Check a stateful coordination protocol with an executable model of competing actors, interruption, and recovery; report bounded safety evidence, progress assumptions, and counterexamples.
Resolve technical options, interfaces, impact, migration, and validation for a requested change before implementation commitments.
Reproduce a reported software failure, distinguish competing causes with decisive probes, and verify an authorized repair against the original symptom.
Choose the smallest sufficient engineering procedure for a requested codebase task and continue to its authorized endpoint.
Complete accepted software intent in small behaviorally useful slices, preserving concurrent work and validating the requested result.
Diagnose a software performance limiter and evaluate a bounded improvement using equivalent work, comparable measurements, and correctness evidence.
Audit testing confidence or create, qualify and refresh consumer verification instructions, separating product, test, harness and environment gaps.
Set up Product Development's packaged Codex agents and guide native hook review, installation, upgrades and verification.
Clarify domain vocabulary, identity, state, invariants, and boundary contracts when ambiguous concepts or repeated rules complicate software design.
Refactor a substantial code or test cluster in evidence-backed phases while preserving caller contracts, behavioral confidence, and relevant runtime outputs.
Run a bounded disposable experiment to answer a specific technical or interaction uncertainty with explicit decision criteria.
Reconcile selected project solution records with current evidence, preserving accepted policy, historical reasoning, and unverifiable claims.
Exercise a planned schema, API, persisted-state, or component migration through supported intermediate versions, interruption, and recovery in an isolated environment before rollout.
Review an exact code candidate for supported defects, regressions, requirement gaps, and meaningful missing evidence using relevant lenses.
Design specification/regression tests or explicitly labeled characterization through a faithful boundary, with meaningful oracles, isolated fixtures and observable results.
Explain a bounded code path or architectural decision, or assess requested maintainability questions from current source without changing product behavior.
Exercise an actual user-facing product path and preserve evidence of its observable outcome, failure, or unavailable prerequisites.