translate
tuesd4y/agent-skills · skills.sh
Open source Repository Open in the app JSON README (API)
About
Skill publicada por tuesd4y/agent-skills no skills.sh. Instale com: npx skills add tuesd4y/agent-skills@translate
Details
- Kind
- Agent skills
- Publisher
- tuesd4y
- Origin
- skillssh
- Category
- ferramentas
- Last push
- 2026-10-06T08:17:15Z
- Repository state
- ativo
- Language
- Shell
- Added
- 2026-10-07 05:26:01
- Updated
- 2026-10-07 05:26:01
- Origin id
tuesd4y/agent-skills/translate
README
# Agent Skills
A collection of skills for AI coding agents. Skills are packaged instructions and scripts that extend agent capabilities.
Skills follow the [Agent Skills](https://agentskills.io/) format.
## Available Skills
### translate
Automated translation workflow for Angular i18n XLF files. Finds new translations in changed files, provides context, and generates translations for review.
Use when:
- After extracting new i18n strings (nx run <app>:extract-i18n)
- When `messages.{lang}.xlf` files have `state="new"` entries
- Before committing translation file changes
- Translating to German, French, Spanish, or Hungarian
Workflow:
1. Detects changed `messages.*.xlf` files via git
2. Extracts entries with `<target state="new">`
3. Finds source context (TypeScript `$localize`, HTML `i18n` attributes)
4. Proposes translation with context for review
5. Applies approved translations with `state="ai"` marker
Features:
- Placeholder preservation (interpolations, HTML tags, line breaks)
- Language detection from filename pattern
- Code context lookup for translation IDs
- Common translation patterns for each language
- Batch processing with summaries
### commit
Commit changed files with a minimal, optionally JIRA-prefixed message. Runs on Sonnet in Claude Code.
Use when:
- Committing finished work (`/commit`, optionally with file paths or extra context)
- You want a tight, review-friendly commit message without boilerplate
Workflow:
1. Detects the JIRA issue key from the branch name (falls back to conversation context)
2. Stages only task-related files by explicit path (never `git add -A`)
3. Drafts a minimal imperative commit message (≤ 100 chars, body only when needed)
4. Confirms files and message with the user before committing
5. On commit hook failures: auto-fixes trivial issues (whitespace, formatting, re-staging
hook-modified files) and retries; asks before substantive fixes (lint/type errors,
failing tests, message policy). Never uses `--no-verify`.
### pr
Open a minimal, structured pull request for the current branch.
Use when:
- Opening a PR for finished branch work (`/pr`, optionally with a base branch, `draft`, or extra context)
- You want a short, structured description (bullets over prose) without boilerplate
- Stacking a PR on top of another open PR you branched off
Workflow:
1. Gathers context in one script call: base branch, tracked `gh stack` state and parent PR
candidates, commits and diff vs merge-base, push state, PR templates, existing PR for the branch
2. Targets the parent PR's branch instead of the default base when stacked PRs are enabled
(`git config skills.pr.stacked true`, or `/pr stacked`), then links the PRs into a GitHub
Stack with `gh stack link`
3. Drafts a short imperative title (JIRA-prefixed when the branch names a ticket);
suggests a real `feature/`/`bugfix/` remote branch name when the local name is temporary
4. Fills the repo's PR template if one exists (keeps all sections and checklists);
otherwise uses a minimal summary + Changes + Notes structure
5. Confirms title and body with the user before pushing the branch and creating the PR
6. Updates the existing PR instead if the branch already has one
Requirements:
- `gh` CLI authenticated for the repo's remote
- For stacked PRs: the `gh stack` extension (`gh extension install github/gh-stack`) — the skill
offers to install it, and falls back to a plain PR against the parent branch if you decline
### work
Start working on a Jira ticket end-to-end. Fetches issue details via Atlassian MCP, creates a worktree with a properly-named branch, gathers context, and enters planning mode.
Use when:
- Starting work on a Jira ticket (`/work BRO-67`)
- Bootstrapping a feature, bugfix, or hotfix from a ticket
Workflow:
1. Parses the Jira issue key from arguments (supports bare keys, URLs, `--hotfix` flag)
2. Fetches issue details, description, and linked issues via Atlassian MCP
3. Creates a git worktree with a conventionally-named branch (`feature/`, `bugfix/`, or `hotfix/`)
4. Transitions the issue to "In Progress"
5. Gathers context: parent epic, linked issues, Confluence pages, sibling tasks
6. Presents a structured summary with open questions
7. Clarifies ambiguities, then enters plan mode
Requirements:
- Atlassian MCP plugin configured with Jira access
### explain-diff
Create a rich, interactive explanation of a diff, commit, branch, or PR as a single self-contained HTML file.
Inspired by [Geoffrey Litt's Explain Diff prompt](https://gist.github.com/geoffreylitt/a29df1b5f9865506e8952488eac3d524).
Invoke explicitly:
- Claude Code: `/explain-diff <change>`
- Codex: `$explain-diff <change>`
- This skill never activates automatically.
Output:
- Beginner and focused background, core intuition with concrete examples and diagrams, and a grouped code walkthrough
- Five interactive multiple-choice questions with explanatory feedback
- Responsive HTML with inline CSS and JavaScript, saved outside the repo as `/tmp/YYYY-MM-DD-explanation-<slug>.html`
## Installation
```bash
npx skills add tuesd4y/agent-skills
```
## Usage
Skills are available once installed. Automatic activation depends on each skill's invocation policy; `explain-diff` requires explicit invocation.
**Examples:**
```
Translate the new translations
```
Or you can manually call the skill e.g. in claude code
`/translate`
## Skill Structure
Each skill contains:
- `SKILL.md` - Instructions for the agent
- `scripts/` - Helper scripts for automation (optional)
- `references/` - Supporting documentation (optional)
## License
MIT