subagent-development

Use when a task can be isolated into a fresh-context subagent or parallelized safely, and its result needs review against the original requirements.

Skill by leo110047

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/leo110047-subagent-development

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 when a task can be isolated into a fresh-context subagent or parallelized safely, and its result needs review against the original requirements. Best fit for self-contained work, not shared-context debugging.

Use cases: productivity, agent-building

Reviews

Write a review

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/subagent-development

What's inside

Skill name: subagent-development
Description: Use when a task can be isolated into a fresh-context subagent or parallelized safely, and its result needs review against the original requirements. Best fit for self-contained work, not shared-context debugging.
License: MIT
Instructions: 47 lines in SKILL.md · view on GitHub

Agent Skills spec issues

allowed-tools: must be a space-separated string

Files (1)

SKILL.md

SKILL.md (first lines)

# Subagent Development

Delegate only when the host surface and task policy allow it and the work is
independent, bounded, and specifiable without hidden conversation context.

## Use When

- The subtask has a clear owner, files, and expected output.
- It runs without touching shared state or blocking the next local step.
- Fresh context improves focus, review quality, or search breadth.
- The result is verifiable from files, commands, or tests.

## Avoid When

- The task is shared-context debugging, unresolved architecture, or an active
  incident whose next step depends on the current investigation.
- Another worker edits the same files.
- Safe execution needs long unstated context.
- No measurable speed or quality benefit exists.

## Dispatch Contract

The prompt contains only what the worker needs:

- Goal: one sentence.
- Scope: exact files, module, or responsibility.
- Context: files to read and patterns to follow.
- Deliverable: what to change or report.
- Boundaries: files, side effects, and scope expansions to avoid.
- Verification: commands or evidence required before reporting completion.

Workers share the codebase with others: they must not revert others' changes and
must adapt to concurrent work.

## Review Contract

A completion report is not evidence. Before accepting it:

- Read every changed file or cited evidence path.
- Map the result to the original requirements.
- Rerun or inspect the relevant verification.
- Check for unrequested scope expansion.

When the result is incomplete, identify what the prompt omitted before
re-dispatching.

From leo110047/goldband, MIT.

Signals

Stars: 0
Forks: 0
Open issues: 5
Last commit (this folder): Oct 14, 2026 (15 days ago)

From GitHub (fetched Oct 24, 2026).

Details

License: MIT
Repository: github.com/leo110047/goldband
First seen: Oct 24, 2026
Last checked: Oct 24, 2026
Source: GitHub github-graphql

More from leo110047

Cite this page

subagent-development skill. Agent Mart. https://agentmart.dev/skills/leo110047-subagent-development (last updated 2026-10-24)