Agent skill

Implement Plan

by jackfranklin in jackfranklin/dotfiles

Implements an explicitly approved, scoped code change safely and autonomously.

MITAuto-check passedDevelopment

Install Implement Plan

skills CLI
$ npx skills add jackfranklin/dotfiles --skill implement-plan -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI
$ gh skill install jackfranklin/dotfiles implement-plan --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy
$ git clone --depth 1 https://github.com/jackfranklin/dotfiles.git skills-src && mkdir -p .claude/skills && cp -r skills-src/claude/skills/implement-plan .claude/skills/implement-plan && rm -rf skills-src

Use ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.

Claude Code skills documentation · loads skills from .claude/skills/

Facts

Skill name
implement-plan
GitHub stars
255
Token cost
~1.9k tokens
SKILL.md length
1,073 words
Files
1
Skills in repo
19
Repo updated
First seen
Licence
MIT

At a glance

Implements an explicitly approved, scoped code change safely and autonomously.

  • Works in 7 steps: Preflight → Branch safely → Confirm the test plan → …
  • Transitioning from an agreed plan to source
  • SKILL.md covers Non-negotiable rules, Working mode and Workflow
  • Calls git

What it does

Implement Plan is an agent skill from jackfranklin/dotfiles. Implements an explicitly approved, scoped code change safely and autonomously. Use when transitioning from an agreed plan to source or test changes; do not use for exploration or unresolved design decisions.

Its SKILL.md is about 1.9k tokens, which your agent loads only when the skill is triggered. It is a single SKILL.md file with no bundled scripts.

It sits in Development, covering Architecture decision records. The repository describes itself as: My dotfiles for my dev environment, compromising of tmux, vim, zsh and git. The licence is MIT.

When your agent uses it

  • Transitioning from an agreed plan to source
  • Do not use for exploration
  • Unresolved design decisions

Example prompts

  • “Use the implement-plan skill to implement an explicitly approved, scoped code change safely and autonomously”
  • “/implement-plan”

Workflow steps

7 steps, taken from the step headings in SKILL.md.

  1. Preflight
  2. Branch safely
  3. Confirm the test plan
  4. Validate continuously
  5. Commit logical, verified increments
  6. Push and open a pull request
  7. Report and hand off

What it can do on your machine

Read from SKILL.md and the folder at commit 7ec4998. It shows what the files ask for, not the result of running them.

  • Tool permissions

    Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

    From allowed-tools in the SKILL.md frontmatter.

  • Runs code

    Shell commands in SKILL.md call:

    • git

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    No URLs in SKILL.md. Its commands use git, which can reach the network depending on how they are called.

    From URLs in SKILL.md, links to its own repository left out.

  • Credentials

    Names no API keys, tokens, secrets or passwords.

    From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Implement Plan loads about 1.9k tokens when it runs. Until then it costs about 56 tokens; SKILL.md has 1,073 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~56
When it runs · the whole SKILL.md, loaded when a task matches
~1.9k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check passed

The automated check found no risky patterns in SKILL.md.

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from jackfranklin/dotfiles at commit 7ec4998, republished under its MIT licence (© jackfranklin). 1,073 words, ~1,893 tokens.

Download SKILL.mdSave it as .claude/skills/implement-plan/SKILL.md (or your agent's skills folder).
name
implement-plan
description
Implements an explicitly approved, scoped code change safely and autonomously. Use when transitioning from an agreed plan to source or test changes; do not use for exploration or unresolved design decisions.

Implement an Approved Plan

Use this skill only to implement a defined change. Do not start source-code changes until the plan is clear and the user has authorized implementation.

Non-negotiable rules

  1. Implement only a clear plan. Read the plan, relevant source, tests, and repository instructions. Confirm the intended behavior, scope, constraints, affected files, and acceptance criteria. If any material detail is ambiguous, stale, contradictory, or missing, stop before editing and explain the gap with a focused question. Do not fill in requirements from guesswork.
  2. Resolve ordinary implementation friction autonomously. Continue through incomplete intermediate states, type errors caused by the current refactor, test failures that can be investigated from repository evidence, and fixture or test-harness changes required by the approved plan. Stop only when resolving the issue requires a material product decision, changes the approved scope, needs unavailable access or secrets, risks data loss, or cannot be resolved from repository evidence. When stopping is necessary, explain the concrete blocker and ask one focused question; do not merely announce a pause.
  3. Use an implementation branch by default; honor explicit authorization to work on main. Before any source or test edit, inspect the Git state. Start from a clean worktree and create/switch to a new, clearly named feature branch unless the user explicitly authorizes implementation on main for the current task. That authorization permits source edits and commits directly on main; do not require a branch or ask again. If the worktree is dirty, the current branch is not suitable, or branch creation would discard/conflict with work, stop and ask for guidance.
  4. Test thoroughly by default. Add or update focused unit tests unless the user explicitly says not to. Cover required happy paths, meaningful boundaries, failure/empty states, and behavior likely to regress from the change. Follow repository test conventions; do not add redundant or speculative tests merely to increase the count.
  5. Prefer the simplest clear implementation. Make the narrowest change that meets the approved plan. Prefer direct, readable code over a new abstraction, layer, configuration option, dependency, state model, or extension point. Introduce one only when a current requirement, two real current use cases, or an established repository convention justifies it. Do not refactor nearby code merely to make the change feel cleaner.

Working mode

During an approved implementation, keep taking the next concrete investigation, edit, or verification step. Do not end a response merely to provide a progress update, announce the next step, or ask for confirmation already supplied. Return to the user only with a completed result or a genuine blocker under rule 2.

Workflow

1. Preflight
  • Read the nearest applicable AGENTS.md, contributor guidance, and relevant task/issue/PR material.
  • Inspect the current implementation, related tests, existing utilities, and recent relevant changes before designing new code.
  • Check git status --short and the current branch.
  • Establish the implementation understanding, including non-goals and acceptance criteria. Do this silently unless a material ambiguity requires a focused question.
2. Branch safely
  • If currently on main with a clean worktree, create and switch to a descriptive branch (for example, feature/issue-123-short-description) unless the user explicitly authorized implementation on main for this task. When such authorization was given, work on main without requesting branch confirmation.
  • If already on an explicitly designated, clean implementation branch, confirm it is appropriate before using it.
  • Never use forceful Git operations, overwrite unrelated changes, or alter another branch's history without explicit user authorization.
3. Confirm the test plan

Before writing tests, derive the specific test cases to add or change from the approved plan and repository guidance. Ask the user only when repository guidance explicitly requires test-plan approval or the expected behavior is materially ambiguous.

Implement tests and the smallest clear production change that satisfies the plan. Reuse an existing abstraction when it fits; otherwise prefer local, direct code to creating a new general abstraction for one use case. Keep unrelated cleanup out of the change.

Show full SKILL.md (427 more words)Show less
4. Validate continuously
  • Run the repository-required typecheck/lint command before tests when instructed by project guidance.
  • Run focused tests while implementing, then the required broader verification once the change is complete.
  • Investigate and fix failures attributable to the in-progress implementation. Do not weaken tests to hide a defect. Stop only for a failure that is pre-existing, unrelated, or cannot be resolved from repository evidence; report the evidence and its impact.
  • Check formatting and inspect the final diff for unintended changes, missing tests, debug code, deviations from the approved scope, and indirection that can be removed without losing a current requirement.
5. Commit logical, verified increments

Commit each independently coherent, verified unit of work as soon as it is complete—for example, a focused behavior change with its tests, followed by a separate integration or documentation change. Before each commit, inspect the relevant status and diff, and commit only files within the approved scope with a concise message describing that unit.

Do not create artificial, incomplete, or WIP commits. Keep production code and the tests that establish its behavior together when they form one logical unit. For small or indivisible work, one final commit is correct; use multiple commits only when the implementation naturally proceeds in separable, reviewable stages.

Do not amend, force-push, or include unrelated user changes.

6. Push and open a pull request

After the full change passes its required verification, inspect the final Git status and diff. Commit any remaining complete logical unit; do not manufacture a final commit when none remains.

Inspect the branch's configured remote URL. If it is hosted on GitHub, push the implementation branch and create a pull request against the repository's default branch with gh. Use the branch's commit series to provide a clear PR title and description. If a pull request for that branch already exists, do not create a duplicate; report its URL instead. If the remote is not GitHub, do not push or create a pull request.

Do not merge branches or modify issue state unless the user explicitly requests it. Never force-push or perform destructive remote operations without explicit user authorization.

7. Report and hand off

Report:

  • branch name and commit SHA(s), in commit order;
  • files changed and the behavior implemented;
  • tests and verification commands run, with results;
  • pull request URL when one was created or already existed, or that no GitHub remote was configured;
  • anything not run or any remaining manual checks;
  • any intentional deviations from the plan, which require prior user approval;
  • complexity deliberately avoided, and any new abstraction or moving part with its present-day justification.

© jackfranklin, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

Just SKILL.md in claude/skills/implement-plan of jackfranklin/dotfiles.

Open the folder on GitHubat commit 7ec4998

Compare with similar skills

Implement Plan next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

Implement Plan compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Implement Plan this skilljackfranklin/dotfiles255—~1.9kAutomated safety check: PassMIT
Grill With Docsmeain/dotfiles28521 repos~875Automated safety check: PassMIT
Development Workflowrunceel/ReactiveProperty944—~1.4kAutomated safety check: PassMIT
Grill With Docsayoubben18/ab-method192—~2.1kAutomated safety check: PassMIT
Architecture Guidelineseser/stack128—~516Automated safety check: PassCustom licence
Dsh Code ReviewZhou-Yujing114514/deepseek-harness-linux116—~2.1kAutomated safety check: PassMIT

Similar skills

  • Grill With Docs

    meain/dotfiles

    Grilling session that challenges your plan against the existing domain model, sharpens terminology, and updates documentation (CONTEXT.md, ADRs) inline as decisions crystallise.

    285 GitHub starsUsed in 21 repos~875 tokens
    DevelopmentAuto-check passed
  • Development Workflow

    runceel/ReactiveProperty

    ReactiveProperty repository development policy. An agent skill from runceel/ReactiveProperty.

    944 GitHub stars~1.4k tokensUpdated 1 mo ago
    DevelopmentAuto-check passed
  • Grill With Docs

    ayoubben18/ab-method

    Grilling session that challenges your plan against the existing domain model, sharpens terminology, and updates documentation (CONTEXT.md, ADRs) inline as decisions crystallise.

    192 GitHub stars~2.1k tokensUpdated 6 days ago
    DevelopmentAuto-check passed
  • System design in eserstack: double-layered hexagonal architecture, explicit adapter composition, public API and CLI surface, request metadata, testing strategy, ADRs.

    128 GitHub stars~516 tokensUpdated 4 days ago
    DevelopmentAuto-check passed
  • Dsh Code Review

    Zhou-Yujing114514/deepseek-harness-linux

    A skill your agent uses when reviewing a pull request in the deepseek-harness repo — orients the reviewer to this codebase's standards (AGENTS.md conventions, defensive patterns, ADRs, quality…

    116 GitHub stars~2.1k tokensUpdated yesterday
    DevelopmentAuto-check passed
  • Technical Design Doc Creator

    tech-leads-club/agent-skills

    Creates comprehensive Technical Design Documents (TDD) with mandatory and optional sections through interactive discovery.

    7k GitHub stars~13k tokensUpdated 17 days ago
    DevelopmentAuto-check passed

More from jackfranklin/dotfiles

All 19 skills in this repo
  • GitHub Code Review

    jackfranklin/dotfiles

    Perform a thorough, read-only review of one GitHub pull request.

    255 GitHub stars~1.8k tokensUpdated today
    Auto-check passed
  • Adr

    jackfranklin/dotfiles

    Capture an Architecture Decision Record (ADR) for a significant decision made in the current project.

    255 GitHub stars~962 tokensUpdated today
    Auto-check passed
  • Jack References

    jackfranklin/dotfiles

    Manage Jack's personal technical reference library at ~/git/references.

    255 GitHub stars~1.2k tokensUpdated today
    Auto-check passed
  • Later

    jackfranklin/dotfiles

    Log items to come back to later — bugs found mid-task, feature ideas, project feedback — as GitHub Issues.

    255 GitHub stars~767 tokensUpdated today
    Auto-check passed
  • New Deno App

    jackfranklin/dotfiles

    Scaffold a new Deno 2 + Hono + Deno KV + Eta + HTMX app with password auth and PWA support.

    255 GitHub stars~3.6k tokensUpdated today
    Auto-check: notes
  • Resolve Merge Conflict

    jackfranklin/dotfiles

    A skill your agent uses when you need to resolve an in-progress git merge/rebase conflict.

    255 GitHub starsUsed in 21 repos~427 tokens
    Auto-check passed

Questions about Implement Plan

What does Implement Plan do?

Implements an explicitly approved, scoped code change safely and autonomously. Implement Plan is an agent skill from jackfranklin/dotfiles. Implements an explicitly approved, scoped code change safely and autonomously.

When should I use Implement Plan?

Implement Plan fits situations like: transitioning from an agreed plan to source; do not use for exploration; unresolved design decisions.

How do I install Implement Plan in Claude Code?

Run `npx skills add jackfranklin/dotfiles --skill implement-plan -a claude-code`. Or copy the skill folder (claude/skills/implement-plan in jackfranklin/dotfiles) into .claude/skills/implement-plan in your project. Claude Code loads it when a task matches its description.

How do I install Implement Plan in Codex?

Run `npx skills add jackfranklin/dotfiles --skill implement-plan -a codex`. Or copy the skill folder (claude/skills/implement-plan in jackfranklin/dotfiles) into .agents/skills/implement-plan in your project. Codex loads it when a task matches its description.

Can I use Implement Plan in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add jackfranklin/dotfiles --skill implement-plan -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/implement-plan, .gemini/skills/implement-plan, .github/skills/implement-plan and .opencode/skills/implement-plan in your project.

What does Implement Plan need to run?

Going by SKILL.md and its folder, Implement Plan needs the command-line tools its instructions call (git).

Does Implement Plan access the network?

SKILL.md contains no URLs. Its commands use git, which can reach the network depending on how they are called. This is read from the text; nothing was executed.

Is Implement Plan safe to install?

Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.

What licence does Implement Plan use?

Implement Plan is published under the MIT licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Implement Plan use?

About 1.9k tokens (SKILL.md is roughly 7.6k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.

What are the alternatives to Implement Plan?

Skills that share tags, products or a category with Implement Plan: Grill With Docs (meain/dotfiles, 285 stars), Development Workflow (runceel/ReactiveProperty, 944 stars), Grill With Docs (ayoubben18/ab-method, 192 stars) and Architecture Guidelines (eser/stack, 128 stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Implement Plan?

jackfranklin (a GitHub user) maintains it in jackfranklin/dotfiles, which has 255 GitHub stars. The repository holds 19 skills in this directory. The repository was last updated on October 7, 2026.

Source: jackfranklin/dotfiles on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.