native-tools-enforcer

One-time setup skill that explores a project's codebase, drafts project-specific extension content conversationally, and writes the per-skill extension files for the software-writer plugin's writing-code, writing-tests, and writing-docs skills (delivery on Claude Code ships with software-writer 2.x; Codex delivery via AGENTS.override.md). Includes a re-sync mode that audits existing extension files for drift and migrates v1 hook-contexts setups. Long plugin and skill names are intentional to prevent accidental invocation.

Three universal, opinionated skills for writing software: writing-code (API-consultation gate, dependency shape, stack footgun check, two-tier comment policy), writing-tests (classicist behavior-first workflow with stack-specific idioms), and writing-docs (single-owner surfaces, pointer-file discipline, anti-slop validation via the human-author:ai-slop-writing-fixer subagent). Each skill is extendable at every step via Pre-Step-N / Post-Step-N positions and via recognized named configuration values; the plugin ships hooks that deliver a project's extension files (.claude/extensions/software-writer/) wrapped in a structural envelope.

Orchestrate implementation and review work through dispatched workers — Codex CLI (GPT-5.6) and Claude subagents — with a task-strategy workflow, model routing, fenced write scopes, and independent two-worker verification

Enforces Sonnet model for the Plan subagent. Ensures thorough architectural reasoning during plan mode.

Debug AI tooling behavior by analyzing output against source instructions and enabling honest self-diagnosis.