Skip to content
Agentmart
Submit a tool

implement

Implement a piece of work from a spec, bd ticket, or direct request.

Skill by baranbingol1

IndexedMIT
Write a review

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/baranbingol1-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.

We use your email only for this newsletter. Privacy

About

Implement a piece of work from a spec, bd ticket, or direct request. Use when a ticket is ready to build, the user says implement, build, or add a feature, or asks for a refactor. Handles ticket resolution, the build loop, verification, and closing the ticket with evidence.

No 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
Implement a piece of work from a spec, bd ticket, or direct request. Use when a ticket is ready to build, the user says implement, build, or add a feature, or asks for a refactor. Handles ticket resolution, the build loop, verification, and closing the ticket with evidence.
License
MIT
Instructions
56 lines in SKILL.md · view on GitHub

Files (3)

  • SKILL.md
  • feature.md · references/
  • refactoring.md · references/

SKILL.md (first lines)

# implement

Build the work described by a bd ticket, a spec, or the user's request. You own the
design and the verification; delegate bulk edits when the work is long-running
([guard-the-context-window](../guard-the-context-window/SKILL.md)).

## 1. Resolve the work

- A bd id: `bd show <id>`. Fuzzy words ("the auth ticket"): resolve via `bd list` +
  title match; more than one plausible match, name them and ask.
- Mark it in progress (`bd update <id> --status in_progress`). Read the acceptance
  criteria — they are the completion contract.
- No ticket (small direct request): proceed; suggest
  [to-spec-bd](../to-spec-bd/SKILL.md) only when the request is genuinely big.

## 2. Load the playbook

Match the task shape and read the matching playbook, then copy its steps into your todo
list verbatim. A step you choose not to do stays in the list with a one-line
`skip: <reason>`; skipping silently is not allowed.

- New behavior → [references/feature.md](references/feature.md)
- Structure changes, behavior stays → [references/refactoring.md](references/refactoring.md)
- Something is broken → hand off to the [diagnose](../diagnose/SKILL.md) skill instead.

## 3. Build

- Walk the [lazy-ladder](../lazy-ladder/SKILL.md) before writing anything.
- Test-first at the pre-agreed seams via [tdd](../tdd/SKILL.md); the spec's Testing
  Decisions name the seams. No spec: agree the seams before the first test.
- Typecheck regularly, run single test files regularly, the full suite once at the end.
- Comments and docs are welcome where they carry constraints the code can't; never
  strip existing ones.
- Integrate new requirements as if they had been foundational from day one, not bolted
  on; propagate through types, docs, and examples.
- Stateful operations answer "what happens if this runs twice? crashed halfway?" —
  add the reconciliation step when the answer is "depends what was left behind".

## 4. Verify and deliver

- Prove it on the real artifact ([prove-it-works](../prove-it-works/SKILL.md)); if the
  repo has a generated verify skill, run it. Every acceptance criterion checked, each
  against the real path, outcomes noted — not command names.
- Commits are small, ordered, and tell the story
  ([sequence-verifiable-units](../sequence-verifiable-units/SKILL.md)).
- PR descriptions use these sections, dropping empty ones: **Why** (intent, why this
  approach), **Scope** (facts from the diff, real symbols and paths), **Tradeoffs**
  (real choices only), **Blast Radius** (who and what it touches, why it's safe or
  risky), **Verification** (how each check ran and its outcome). Attach a screenshot or
  recording when it proves a claim.
- Close the ticket with the evidence note: `bd close <id>` after updating it with what
  shipped and how it was verified. Pausing instead: leave a resume note on the issue
  (progress, verified-so-far, next action) and a `wip:` commit — never stop mid-edit in
  a broken state.

From baranbingol1/bbstack, MIT.

More from baranbingol1/bbstack

Signals

Stars
0
Forks
0
Open issues
0
Last commit (this folder)
Aug 28, 2026 (2 months ago)

From GitHub (fetched Oct 24, 2026).

Details

License
MIT
First seen
Oct 24, 2026
Last checked
Oct 24, 2026

More from baranbingol1

Cite this page

implement skill. Agent Mart. https://agentmart.dev/skills/baranbingol1-implement (last updated 2026-10-24)