By dr-data
Jules API client skills (v1alpha) — setup, sources, start, sessions, activities. Toggle the whole namespace on/off as one unit.
Inspect activity timeline of a Jules session — agent/user messages, generated plan, plan-approved markers, progress updates, and produced artifacts (code patches as git unidiff, bash command output, media). Use when the user asks "what is Jules doing", wants to read the proposed plan before approving, wants to see the diff Jules generated, or needs the failure reason for a FAILED session.
Manage existing Jules sessions — list all, get one by id, send a follow-up message, or approve a pending plan. Use when the user wants to check session status, see all running/past Jules sessions, send instructions to a paused session ("tell Jules to also update the README"), or approve a plan that is waiting in AWAITING_PLAN_APPROVAL. For activity-level detail (plan steps, code patches, bash output) use /jules:activities instead.
One-time setup for Jules API access. Stores the user's API key at ~/.config/jules/api_key with chmod 600 so other /jules:* skills can read it without loading the key into context. Use when the user runs /jules:setup, when any /jules:* command reports a missing key, or when the user asks how to configure Jules API credentials.
List or fetch Jules sources (connected GitHub repos) via the Jules v1alpha API. Use when the user wants to see which repos Jules can act on, look up a source ID before starting a session, check default branch / branch list of a connected repo, or verify the Jules GitHub app is installed on a repo.
Create a new Jules session via POST /v1alpha/sessions. Use when the user wants to kick off a Jules coding task on a connected GitHub repo — e.g. "have Jules fix this bug", "start a Jules session on owner/repo", "ask Jules to refactor X", or wants the session to auto-open a PR. Prompts user for the repo (source) and branch via /jules:sources if not provided, supports plan-approval gating and AUTO_CREATE_PR.
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
A Claude Code plugin that wraps the Jules API (v1alpha) as five focused slash commands. Lets you drive Jules — Google's autonomous coding agent — without leaving your terminal.
/jules:setup # one-time API-key setup
/jules:sources # list connected GitHub repos
/jules:start # create a new Jules session (prompt + repo + branch)
/jules:sessions # list / get / sendMessage / approvePlan
/jules:activities # plan, diffs, bash output, failure reasons
Toggle the whole jules:* namespace on or off as one unit via /plugin enable|disable. Disabling reclaims the description tokens of all five commands at once.
In a Claude Code session:
/plugin marketplace add dr-data/jules-skills
/plugin install jules@jules-skills
That's it. The five /jules:* commands are now available.
To verify they loaded, type / and look for jules:setup, jules:sources, jules:start, jules:sessions, jules:activities.
The plugin never stores the key, never reads it into Claude's context.
/jules:setup and paste it when asked, or set it up yourself:
mkdir -p ~/.config/jules && chmod 700 ~/.config/jules
printf %s 'YOUR_KEY' > ~/.config/jules/api_key
chmod 600 ~/.config/jules/api_key
JULES_API_KEY=... in your shell — the plugin checks the file first, falls back to the env var.Each command's curl invocation pulls the key at runtime with:
H="X-Goog-Api-Key: $(cat ~/.config/jules/api_key 2>/dev/null || printf %s "$JULES_API_KEY")"
So the key lives on disk only, never inside the loaded skill markdown.
/jules:sources # find your repo's source id
/jules:start # describe the task; pick source + branch
# → returns a session id and a jules.google.com URL
/jules:activities <SID> # watch progress; see the plan
/jules:sessions approve <SID> # if requirePlanApproval=true
/jules:activities <SID> # read generated diff / bash output
/jules:sessions send <SID> "also update the README" # follow-up
For sessions started with automationMode: AUTO_CREATE_PR, Jules opens the PR itself; the URL appears in outputs[].pullRequest.url on the session resource.
/plugin disable jules@jules-skills
/plugin enable jules@jules-skills
Or use the /plugin UI.
/plugin uninstall jules@jules-skills
/plugin marketplace remove jules-skills
The on-disk key at ~/.config/jules/api_key is not touched by uninstall — remove it manually with rm ~/.config/jules/api_key if you want it gone.
jules-skills/
├── .claude-plugin/
│ ├── marketplace.json # marketplace name: jules-skills
│ └── plugin.json # plugin name: jules
├── commands/
│ ├── setup.md → /jules:setup
│ ├── sources.md → /jules:sources
│ ├── start.md → /jules:start
│ ├── sessions.md → /jules:sessions
│ └── activities.md → /jules:activities
└── README.md
Each commands/*.md is a Claude Code slash-command spec: YAML frontmatter (name, description) plus the procedural body Claude reads when the command fires.
| Command | Endpoint |
|---|---|
/jules:sources | GET /v1alpha/sources, GET /v1alpha/sources/{id} |
/jules:start | POST /v1alpha/sessions |
/jules:sessions (list/get) | GET /v1alpha/sessions, GET /v1alpha/sessions/{id} |
/jules:sessions send | POST /v1alpha/sessions/{id}:sendMessage |
/jules:sessions approve | POST /v1alpha/sessions/{id}:approvePlan |
/jules:activities | GET /v1alpha/sessions/{id}/activities[/{aid}] |
Schemas (Session, SourceContext, Activity, Plan, GitPatch, etc.) are summarised inline in each command file — sourced from the official REST reference.
The Jules API is in alpha; field names may shift. If a request fails after the API moves, re-fetch the reference page and update the relevant commands/*.md.
MIT. See plugin.json.
PRs welcome. Keep skill bodies focused: only what Claude can't already infer. No real API keys, no PII, no absolute home-directory paths in committed files.
npx claudepluginhub dr-data/jules-skills --plugin julesComprehensive feature development workflow with specialized agents for codebase exploration, architecture design, and quality review
Reliable automation, in-depth debugging, and performance analysis in Chrome using Chrome DevTools and Puppeteer
Harness-native ECC operator layer - 67 agents, 271 skills, 92 legacy command shims, reusable hooks, rules, selective install profiles, and production-ready workflows for Claude Code, Codex, OpenCode, Cursor, and related agent harnesses
Real-time statusline HUD for Claude Code - context health, tool activity, agent tracking, and todo progress
Comprehensive skill pack with 66 specialized skills for full-stack developers: 12 language experts (Python, TypeScript, Go, Rust, C++, Swift, Kotlin, C#, PHP, Java, SQL, JavaScript), 10 backend frameworks, 6 frontend/mobile, plus infrastructure, DevOps, security, and testing. Features progressive disclosure architecture for 50% faster loading.
Tools to maintain and improve CLAUDE.md files - audit quality, capture session learnings, and keep project memory current.