implement
Build the work a Task ticket describes — claim it, work it on its own branch, prove it with the verify command, and hand back a reviewable branch.
Skill by FastTrackStudios
Install
Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.
npx agentmarket install skill/fasttrackstudios-implement
Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.
Get the weekly diff
The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.
About
Use cases: productivity, agent-building
Reviews
Write a reviewNo reviews yet. Installed it? Reviews come from verified installers only. How reviews work
Permissions
What this skill can do and what it asks for, from its SKILL.md. Checked Oct 25, 2026.
- Tools
- No tools found
- Secrets
- None asked for
- Install scripts
- None found
Network hosts it talks to 1
codeload.github.com· found in install command
Files it touches
~/.claude/skills/implement
What's inside
- Skill name
implement- Description
- Build the work a Task ticket describes — claim it, work it on its own branch, prove it with the verify command, and hand back a reviewable branch.
- License
- Apache-2.0
- Instructions
- 69 lines in SKILL.md · view on GitHub
Files (1)
- SKILL.md
SKILL.md (first lines)
Implement the ticket. Read [../ISSUE-TRACKER.md](../ISSUE-TRACKER.md) first. ## Let the runner do it If a runner is registered for this repo, you do not need to drive any of this by hand: ```bash task runner work --repo <path> --worktree-root <path> \ --agent-cmd 'claude -p --permission-mode acceptEdits' ``` That claims a ticket, cuts a worktree and branch, runs the agent, runs the verify command, commits, and moves the ticket to `needs-review`. `task runner serve` does it continuously. Everything below is what happens inside that loop — and what to do when you are the agent inside it. ## Working a ticket by hand 1. **Claim before touching anything.** `task issue start <id> --as-agent <you>`. Claiming is atomic; if you lose, stop — someone else has it. 2. **Work on the ticket's own branch**, cut from the base. Never move the repository's checkout, and never push. 3. **Use `/tdd`** at the seams the spec agreed. Typecheck as you go, run single test files often, the full suite once at the end. 4. **Run the verify command.** Not a command you think is equivalent — the one the ticket resolves to. Exit zero is the only pass. 5. **`/code-review`** before committing. 6. **Commit to the branch.** The branch is the handback. Nothing merges to a mainline, and nothing is pushed. ## When you are stuck Ask. Do not guess, and do not decide the question is obvious: ```bash task runner ask <ticket> "<the question>" --option a --option b ``` That records the question and flips the ticket to `needs-input`, which takes it out of the runner queue until a human answers. Answering restores `ready-for-agent`. **Never answer your own question.** A human-in-the-loop decision resolves only through the human. If you find yourself reasoning about which answer they would probably pick, that is the moment to ask. ## When you change nothing Say so and make no commit. An empty commit makes the branch lie about what happened. ## When verify fails Leave the worktree. Do not delete evidence to make the next attempt cleaner — the failed tree is what someone will look at. Release the claim so the ticket can be retried; the attempt is kept either way, so "this has died three times on the same command" stays answerable.
From FastTrackStudios/task, Apache-2.0.
More from FastTrackStudios/task
- 0.595score
to-spec
FastTrackStudios
Turn the current conversation into a spec and publish it to Task — no interview, just synthesis of what you have already discussed.
SkillAFreshness: A★ 2Claude Code +9 - 0.595score
triage
FastTrackStudios
Move issues into the agent lane — decide who does it, what proves it done, and which machine can take it.
SkillAFreshness: A★ 2Claude Code +9 - 0.573score
codebase-design
FastTrackStudios
Shared vocabulary for designing deep modules.
SkillAFreshness: A★ 2Claude Code +9 - 0.573score
domain-modeling
FastTrackStudios
Build and sharpen a project's domain model.
SkillAFreshness: A★ 2Claude Code +9 - 0.570score
to-tickets
FastTrackStudios
Break a plan, spec, or the current conversation into tracer-bullet tickets in Task, each declaring its blocking edges, its verify command, and the…
SkillAFreshness: A★ 2Claude Code +9
Signals
- Stars
- 2
- Forks
- 0
- Open issues
- 43
- Last commit (this folder)
- Aug 17, 2026 (2 months ago)
- Last release
- Oct 24, 2026 (5 days ago)
From GitHub (fetched Oct 24, 2026).
Details
- License
- Apache-2.0
- Repository
- github.com/FastTrackStudios/task
- First seen
- Oct 24, 2026
- Last checked
- Oct 24, 2026
- Source
- GitHubgithub-graphql
More from FastTrackStudios
- 0.595score
to-spec
FastTrackStudios
Turn the current conversation into a spec and publish it to Task — no interview, just synthesis of what you have already discussed.
SkillAFreshness: A★ 2Claude Code +9 - 0.595score
triage
FastTrackStudios
Move issues into the agent lane — decide who does it, what proves it done, and which machine can take it.
SkillAFreshness: A★ 2Claude Code +9 - 0.573score
codebase-design
FastTrackStudios
Shared vocabulary for designing deep modules.
SkillAFreshness: A★ 2Claude Code +9 - 0.573score
domain-modeling
FastTrackStudios
Build and sharpen a project's domain model.
SkillAFreshness: A★ 2Claude Code +9 - 0.570score
to-tickets
FastTrackStudios
Break a plan, spec, or the current conversation into tracer-bullet tickets in Task, each declaring its blocking edges, its verify command, and the…
SkillAFreshness: A★ 2Claude Code +9
Cite this page
implement skill. Agent Mart. https://agentmart.dev/skills/fasttrackstudios-implement (last updated 2026-10-24)