Skip to content
Agentmart
Submit a tool

implement

Implement an approved change and complete the named verification and correction loop.

Skill by MarkoArsov

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/markoarsov-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 an approved change and complete the named verification and correction loop.

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
Runs as
Python package

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 an approved change and complete the named verification and correction loop.
License
MIT
Instructions
26 lines in SKILL.md · view on GitHub

Before installing

  • Ships executable scripts

Files (2)

  • SKILL.md
  • resolve.py · scripts/

SKILL.md (first lines)

<!-- resolver:start -->
Run this skill folder's scripts/resolve.py from the active project once. If it returns a different effective skill, follow that file instead. Otherwise continue below. Use the selected package's bin/plovi for commands.
<!-- resolver:end -->

# implement

Read the selected package's references/quality-guidance.md and the plan-selected
project quality sources before changing or judging code. In a detached stage use
only the supplied frozen resources; do not reload live overrides or plugin caches.
In direct use resolve the effective project sources. Report an explicitly selected
missing source rather than silently substituting unrelated guidance.

1. Read the complete plan, project workflow, native instructions, and nearest established implementation. Follow the planned repository membership and path boundaries.

2. Use the current checkout or prepared task worktree. Read-only companion repositories supply context and checks; do not branch or edit them.

3. For a manual bugfix, reproduce the defect, add a regression check, observe its
   behavioral failure, make the minimal fix, rerun it green and run focused checks.
   Record concise before/after evidence. When automation is impractical, explain
   why and use the strongest repeatable alternative. Never call an import/build
   failure behavioral red. Implement the smallest complete change that satisfies the requirements. Handle authorization, data boundaries, failure modes, compatibility, and migrations where the project requires them. Avoid opportunistic refactors.

4. Preserve independent tests. Do not weaken, skip, remove, or rewrite assertion-proven tests to make results green. A wrong test or changed requirement needs a complete plan revision and renewed red evidence.

5. Run each named check in its configured working directory. Diagnose failures, fix the cause, and repeat only relevant checks. Finish with all acceptance mappings evidenced against the current diff.

6. Inspect the final diff for accidental files, credentials, unrelated formatting, and generated artifacts. Follow project rule detectors and scoped references.

7. Do not perform orchestrator-owned commit/push/PR mechanics. In direct use, separate completion from any unrequested delivery action.

8. When blocked by a concrete missing decision or inaccessible dependency, state exactly what is needed and why. Do not invent credentials, silently omit checks, or claim success from an agent narrative.

9. Return complete only when implementation is ready for independent checks; include any material limitations. The orchestrator validates outputs and may return observed failures for correction.

From MarkoArsov/agent-workflow, MIT.

More from MarkoArsov/agent-workflow

Signals

Stars
0
Forks
0
Open issues
0
Last commit (this folder)
Sep 28, 2026 (1 month ago)

From GitHub (fetched Oct 24, 2026).

Details

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

More from MarkoArsov

Cite this page

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