melconcoast
User"code this idea" — turns a plan or idea into a complete, AI-coding-agent-ready docs set (AGENTS.md + linked architecture/decisions/roadmap/product/design docs). Interviews you with grounded recommendations instead of a fixed template. Works with Claude Code, Codex, Cursor, and other AGENTS.md-compatible agents.
Categories
Indexed Skills (5)
execute-plan
Use this skill when the user says things like "execute the plan", "build task 2.1", "start phase 2", "continue building this module", "implement the next task", "work through the plan", or is ready to write code against a plan file that already exists. Not for creating a project's agent docs or its roadmap — use scaffold. Not for cutting a module into phases or re-planning one — use plan-module. Builds a module from the plan file `plan-module` wrote — `docs/guides/feature_[module]_plan.md` — in a strict micro-loop. Implement exactly one task, write the test code its plain-English scenarios describe, run it until green, then update the plan file before touching the next task. Stops at every phase verification gate to report and ask. Keeps closed counts, the Progress Log, Files Modified, and `docs/development-roadmap.md` statuses true as it goes.
plan-module
Use this skill when the user says things like "plan the next module", "plan module 3", "break this module into phases", "create the execution plan for [module]", or is ready to start building a module the roadmap already defines. Not for creating the roadmap or a project's agent docs — use scaffold. Not for building a module that already has a plan file — use execute-plan. Turns one module of an existing development roadmap into a phase-based execution spec an automated coding agent can build from — `docs/guides/feature_[module]_plan.md`, with 2–4 dependency-ordered phases, at most 3–4 tasks each, plain-English test scenarios under every task, and a mandatory verification gate closing each phase. Reads scope, sub-modules, and dependency order out of `docs/development-roadmap.md` rather than inventing them, and confirms the phase cut before writing.
scaffold
Use this skill whenever the user wants to turn a plan/idea into files for Claude Code or another coding agent, says things like "set up AGENTS.md/CLAUDE.md for this", "get this ready for Claude Code", "scaffold the project docs", "turn this plan into context files", "hand this off to a coding agent", or has just finished a substantial planning/design discussion and is about to start building. Also trigger when the user asks how to structure AGENTS.md/CLAUDE.md and wants it actually applied to their project, not just explained in the abstract. Not for planning or building a module from an existing roadmap — use plan-module or execute-plan. Transforms a project plan, idea, or planning conversation into an AI-coding-agent-ready docs set — a lean root context file for the target agent (CLAUDE.md, AGENTS.md, or both), plus linked docs and nested per-subsystem files for monorepos. Interviews to pick the right structure, re-confirms undecided facts, and tracks deferred choices as pending decisions.
test-and-verify
Use this skill whenever tests are to be run or a failure chased down — "run the tests", "run the test suite", "verify this", "check the tests pass", "fix the failing tests", "why is this test failing", or when `execute-plan` needs a task or a `Task X.V` gate verified. Use it even when the request looks like a one-liner — the bounded loop and the honest verdict are the point, not the command. Not for writing a feature or its tests from a plan — use execute-plan. Not for standing up a project's first test runner — use execute-plan. Runs a project's tests, reads the real output rather than the exit code, and fixes what fails — a bounded diagnose-fix-rerun loop that stops after three remediation attempts instead of guessing on, and never weakens a test to reach green. Targets the suite relevant to the work, and adds the type-checker and linter at a phase gate. Reports a plain pass or fail verdict with the exact command it ran, and never edits a plan file — the caller marks the boxes.
code-idea
Transforms a project plan, idea, or planning conversation into an AI-coding-agent-ready docs set — a lean root context file in whichever format the target agent reads natively (CLAUDE.md, AGENTS.md, or both), plus linked docs and nested per-subsystem files for monorepos. Interviews the user to pick the right structure, not one fixed template, re-confirms anything not explicitly decided, and supports deferring a choice as a tracked pending decision. Use this skill whenever the user wants to turn a plan/idea into files for Claude Code or another coding agent, says things like "set up AGENTS.md/CLAUDE.md for this", "get this ready for Claude Code", "scaffold the project docs", "turn this plan into context files", "hand this off to a coding agent", or has just finished a substantial planning/design discussion and is about to start building. Also trigger when the user asks how to structure AGENTS.md/CLAUDE.md and wants it actually applied to their project, not just explained in the abstract.
Bio shown is the top-scored skill's repo description as a fallback — real GitHub bios land in a future update.