foundrylisted

Review and build agentic tooling for correctness, portability, safety, and token cost. Report only genuine problems. Do not nitpick, do not invent issues, and do not rewrite the user's working code unless they ask for the fix to be applied. Read `references/principles.md` first. It carries the five doctrines that govern everything below. Correctness over confidence, precision over zeal, portability first, token budget discipline, lifecycle and clarity. ## What this skill covers Skills, hooks, MCP servers, subagents, slash commands, context engineering, settings, and security. The full agentic stack, not one framework, and across Claude Code, Codex, Gemini CLI, Cursor, Cline, and Copilot. ## Modes The skill accepts an optional argument that narrows the work. - No argument. Review the tooling in scope and report findings. - `build skill|hook|mcp|subagent|command`. Build that artifact to the matching reference, write its test, register it, and confirm it fires. - `review <path>`. Audit existing tooling at that path for the mistakes in `references/anti-patterns.md`. - `harden`. Run a security pass using `references/security.md`. - `port <assistant>`. Map the tooling to another assistant using `references/portability.md`, translating the capability, not only the file. ## Review process Load only the reference files relevant to the artifact in scope. 1. Apply the doctrines in `references/principles.md`. 2. Check skills against `references/skills.md`. 3. Check hooks against