gsd-run

Execute work — full phase, quick task, or trivial inline fix

9 stars

bymlucascosta

View on GitHub Installation ↓

Best use case

gsd-run is best used when you need a repeatable AI agent workflow instead of a one-off prompt.

Execute work — full phase, quick task, or trivial inline fix

Teams using gsd-run should expect a more consistent output, faster repeated execution, less prompt rewriting.

When to use this skill

You want a reusable workflow that can be run more than once with consistent structure.

When not to use this skill

You only need a quick one-off answer and do not need a reusable workflow.
You cannot install or maintain the underlying files, dependencies, or repository context.

Installation

Claude Code / Cursor / Codex

$curl -o ~/.claude/skills/gsd-run/SKILL.md --create-dirs "https://raw.githubusercontent.com/mlucascosta/ia_boilerplate/main/.agents/runtimes/codex/skills/gsd-run/SKILL.md"

Manual Installation

Download SKILL.md from GitHub
Place it in .claude/skills/gsd-run/SKILL.md inside your project
Restart your AI agent — it will auto-discover the skill

How gsd-run Compares

Feature / Agent	gsd-run	Standard Approach
Platform Support	Not specified	Limited / Varies
Context Awareness	High	Baseline
Installation Complexity	Unknown	N/A

Frequently Asked Questions

What does this skill do?

Execute work — full phase, quick task, or trivial inline fix

Where can I find the source code?

You can find the source code on GitHub using the link provided at the top of the page.

SKILL.md Source

<codex_skill_adapter>
## A. Skill Invocation
- Invoked by mentioning `$gsd-SLUG`.
- Treat all user text after `$gsd-SLUG` as `{{GSD_ARGS}}`.

## B. AskUserQuestion → request_user_input Mapping
- `header` → `header`, `question` → `question`
- Execute mode fallback: plain-text numbered list, pick reasonable default

## C. Task() → spawn_agent Mapping
- `Task(subagent_type="X", prompt="Y")` → `spawn_agent(agent_type="X", message="Y")`
- `Task(model="...")` → omit
- Parallel: spawn → collect IDs → `wait(ids)` → `close_agent(id)`
</codex_skill_adapter>

<objective>
Three execution modes:

Default (phase number): Execute all plans with wave-based parallelization.
--quick (or task description): Quick task with GSD guarantees. Add --research or --full for extras.
--fast (trivial): Inline edit, no planning, no subagents.

Routing:
1. --fast → fast workflow
2. --quick or task description → quick workflow
3. --review → review workflow
4. --workspace new|list|remove → workspace workflows
5. --milestone audit|complete → milestone workflows
6. --forensics → forensics workflow
7. Phase number → execute-phase workflow

execute-phase flags: --wave N, --gaps-only, --interactive
quick flags: --research, --full, --discuss
ops flags: --review [phase], --workspace new|list|remove, --milestone audit|complete, --forensics "<issue>"
</objective>

<context>
{{GSD_ARGS}}
</context>

<process>
Parse arguments to determine mode and execute the matched workflow end-to-end.
</process>

<execution_context>
@.agents/workflows/execute-phase.md
@.agents/workflows/quick.md
@.agents/workflows/fast.md
@.agents/workflows/review.md
@.agents/workflows/new-workspace.md
@.agents/workflows/list-workspaces.md
@.agents/workflows/remove-workspace.md
@.agents/workflows/audit-milestone.md
@.agents/workflows/complete-milestone.md
@.agents/workflows/forensics.md
</execution_context>