Back to the catalog

saga

Run an autonomous, spec-driven development "saga" for medium-to-large features using an orchestrator agent and a fleet of worker subagents.

Open source Repository Open in the app JSON README (API)

About

Run an autonomous, spec-driven development "saga" for medium-to-large features using an orchestrator agent and a fleet of worker subagents. Use this skill whenever the user invokes /saga, asks to autonomously build a sizable feature end-to-end with minimal human intervention, wants a comprehensive spec broken into milestones and tasks with airtight validation criteria before parallelized implementation, or wants an orchestrator to delegate implementation to worker agents while preserving its own context window. Trigger on phrases like "run a saga", "autonomously implement this feature", "spec it out then build it with subagents", "orchestrate this big feature end-to-end", or "build this with workers and validate each step". Also use this skill when asked to continue, resume, or pick up an existing saga from its saga directory (e.g. under ~/.sagas).

Details

Kind
Agent skills
Topic
Files & documents
Publisher
warpdotdev
Origin
skillssh
Category
ferramentas
Stars
575
Forks
31
Open pull requests
33
Last push
2026-09-10T22:44:26Z
Repository state
ativo
Language
Python
License
MIT
Added
2026-08-30 15:21:15
Updated
2026-09-10 08:01:57
Origin id
warpdotdev/common-skills/saga

README

# Common Skills

This repository is where common agent skills that should be shared across repositories go.

A skill belongs here when it captures a reusable workflow, convention, or operating procedure that is useful in more than one repository. Repository-specific skills should stay with the repository they apply to unless they can be generalized without losing important context.

## Repository layout

```text
.agents/
  skills/
    <skill-name>/
      SKILL.md
      scripts/      # optional helper scripts
      references/   # optional supporting docs
      assets/       # optional bundled assets
```

Each skill lives in its own directory under `.agents/skills/`. The only required file for a skill is `SKILL.md` with YAML frontmatter containing at least:

- `name`: the kebab-case skill identifier
- `description`: what the skill does and when agents should use it

## Current skills

### Spec workflow

- `write-product-spec` — writes user-facing `PRODUCT.md` specs.
- `write-tech-spec` — writes implementation-oriented `TECH.md` specs.
- `spec-driven-implementation` — guides the full spec-first workflow for substantial features.
- `implement-specs` — implements approved `PRODUCT.md` and `TECH.md` files while keeping specs and code aligned.

### Development workflow

- `create-pr` — guidance for preparing and opening pull requests.
- `write-pr-description` — writes the PR body: template sections, plain-language prose, and reviewer guidance.
- `diagnose-ci-failures` — workflow for inspecting GitHub CI failures and producing a fix plan.
- `fix-errors` — guidance for fixing build, lint, formatting, and test failures.
- `resolve-merge-conflicts` — workflow and helper script for resolving git conflicts with compact context.
- `review-pr` — produces structured PR review feedback from local diff artifacts.
- `check-impl-against-spec` — compares PR implementation changes against provided spec context during review.

### Investigation and decision-making

- `research` — delegates low signal-to-noise-ratio research work to subagents and returns distilled, evidence-backed findings.
- `cross-critique` — sharpens contested decisions by having agents critique one another's independent proposals before synthesis.

### Skill authoring

- `update-skill` — guidance for creating and maintaining skill directories and `SKILL.md` files.
- `skill-doctor` — grades a repo's installed skills by scoring recent local agent conversations, then drafts the skill edits the evidence justifies.

## Adding a shared skill

When adding a skill to this repository:

1. Put it under `.agents/skills/<skill-name>/`.
2. Include a `SKILL.md` with clear frontmatter.
3. Keep the skill focused on a reusable workflow rather than one repository's private details.
4. Move large reference material into `references/` and helper automation into `scripts/`.
5. If copying from another repo, copy first, then generalize in a separate change so the provenance is easy to review.

## Generalizing repository-specific skills

Some skills copied here may still contain repository-specific examples, paths, commands, or assumptions. That is okay during initial migration, but shared skills should eventually be generalized by:

- replacing hard-coded repository names with placeholders or conditional guidance
- separating common workflow guidance from local repository conventions
- moving repo-specific overrides back into the repository that needs them
- keeping descriptions broad enough to trigger in multiple repositories, but specific enough to avoid unrelated tasks

## Usage

Consumers can install the shared skills with the `skills` CLI.

List available skills:

```sh
npx skills@latest add warpdotdev/common-skills --list
```

Install all common skills for Warp globally:

```sh
npx skills@latest add warpdotdev/common-skills --skill '*' --agent warp --global
```

Install one skill:

```sh
npx skills@latest add warpdotdev/common-skills --skill write-tech-spec --agent warp --global
```

Update installed skills later:

```sh
npx skills@latest update --global --agent warp
```

You can also copy or sync selected directories from `.agents/skills/` into a repository's own `.agents/skills/` directory.

Prefer copying only the skills a repository actually needs. If a common skill needs repository-specific behavior, add a small local companion skill in that repository rather than forking the shared skill unless the change is useful everywhere.

More