skills/build-from-exec-plan/SKILL.md
Build a repository-local feature by executing an existing ExecPlan. Use when the user explicitly asks to build a named plan, for example `use the build-from-exec-plan skill to build multi-cluster-profiles`. Resolve the plan by basename under `docs/exec-plans/`, treat the plan file as the source of truth, move the plan to `docs/exec-plans/active/` before implementation if needed, execute it milestone by milestone with verification, create a git commit after each completed milestone, and move the plan to `docs/exec-plans/completed/` when all milestones are finished.
npx skillsauth add neodejack/skills build-from-exec-planInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
Execute an existing repository-local ExecPlan from start to finish. This skill is for implementation, not planning. The plan is the working spec and must stay current as work proceeds.
Resolve the target plan.
use the build-from-exec-plan skill to build <plan_name>.<plan_name> as the basename of a Markdown file under docs/exec-plans/.docs/exec-plans/active/<plan_name>.mddocs/exec-plans/todo/<plan_name>.mddocs/exec-plans/completed/<plan_name>.mdcompleted/, stop and report that the plan is already completed unless the user explicitly asks to reopen it.Read the implementation sources of truth.
Activate the plan lifecycle.
docs/exec-plans/todo/, move it to docs/exec-plans/active/ before implementation starts.active/, keep working there.completed/ until all milestones and acceptance criteria are satisfied.Execute the plan milestone by milestone.
ProgressSurprises & DiscoveriesDecision LogOutcomes & RetrospectiveDecision Log and ensure the rest of the document stays self-consistent.Verify each milestone before declaring it done.
Commit after every completed milestone.
todo/ to active/ happens at the start of the first milestone, include that move in the first milestone commit unless a separate activation commit is necessary to preserve accurate status.Finish and archive the plan.
Outcomes & Retrospective to summarize what was achieved, remaining gaps, and any notable lessons.docs/exec-plans/active/ to docs/exec-plans/completed/.completed/, unless the repository needs a distinct completion commit for clarity.completed/ while milestone work or acceptance remains unfinished.At the end of the task, report:
todo/ to active/ and later to completed/tools
Create, edit, and maintain `justfile` automation for any codebase using the `just` command runner. Trigger whenever a request involves creating a new `justfile`, modifying/updating an existing `justfile`, adding or changing recipes, refactoring `justfile` structure, or automating project commands via `just`.
development
Write documentation for Elixir modules, functions, types, and callbacks following official Elixir conventions. Use when asked to document Elixir code, add @moduledoc/@doc/@typedoc, write doctests, or improve Elixir documentation. Triggers on: document this elixir module, add elixir docs, write moduledoc, add doctests.
tools
Update a Homebrew tap formula to the latest GitHub Release assets using gh CLI, including verifying the tap repo, locating the formula, computing sha256 for release binaries, and committing/pushing changes. Use when asked to bump a Homebrew formula version to the latest release in a personal tap.
development
Bootstrap or improve harness-engineering scaffolding for an existing software repository so agents can work safely and productively. Use when asked to make a repo more agent-friendly, adopt harness engineering, prepare a codebase for Codex or mixed human and agent coding, add or improve repo-local guidance such as `AGENTS.md` or `ARCHITECTURE.md`, establish canonical setup/lint/typecheck/test commands, or audit a repository for missing agent workflows and verification rails.