Skip to content
Back to skills

Mc Clayrune Apis

ASecurity

Use whenever you need to interact with Clayrune local APIs — backlog management, scheduler routines, hivemind orchestration, process registration, terminal pop-out, or any localhost:5199 endpoint. TRIGGER when the user mentions "backlog", "schedule", "hivemind", "register this process", "open a terminal", or refers to Clayrune by name.

  • 16 stars
  • 0 votes
  • 0 copies
  • 6 views
  • Added September 3, 2026
ai-agentsgobashapi

Works with

  • terminal
  • api

Security analysis

A92/100
  • mediumUses curl or wget to download content

Pro shows the line behind each finding and how to fix it

Scanned September 3, 2026

npx -y skills add ronle/clayrune --skill mc-clayrune-apis --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Mc Clayrune Apis?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Mc Clayrune Apis
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/ronle-mc-clayrune-apis/badge)](https://www.skillsdirectory.com/skills/ronle-mc-clayrune-apis)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: mc-clayrune-apis
description: Use whenever you need to interact with Clayrune local APIs — backlog management, scheduler routines, hivemind orchestration, process registration, terminal pop-out, or any localhost:5199 endpoint. TRIGGER when the user mentions "backlog", "schedule", "hivemind", "register this process", "open a terminal", or refers to Clayrune by name.
---

# Clayrune API surface

Clayrune runs locally at **http://localhost:5199**. Every Clayrune-aware operation is a curl call to this base URL.

## Process Registration — MANDATORY for any spawned process

If you spawn a background process, server, bot, or any long-running command, you MUST register it. Unregistered processes cannot be monitored or stopped by the user.

```bash
# 1) Spawn and capture PID
mycommand &
PID=$!

# 2) Register
curl -s -X POST http://localhost:5199/api/processes/register \
  -H "Content-Type: application/json" \
  -d "{\"pid\":$PID,\"name\":\"Short description\",\"project_id\":\"<pid>\",\"command\":\"the command run\"}"
```

PID must be an integer. Do not skip registration.

## Backlog (per-project task list)

When the user says "backlog", "backlog items", "the list" — they mean THIS, not files on disk.

- Read: `GET /api/project/<pid>/backlog`
- Update: `PATCH /api/project/<pid>/backlog/<item_id>` with `{"status":"done"}` — status values: `open`, `in_progress`, `blocked`, `done`, `wontdo`. The first three are LIVE (visible in the tab, counted on the tile); the last two are closed.
- Add note: `POST /api/project/<pid>/backlog/<item_id>/note` with `{"text":"..."}`
- Add item: `POST /api/project/<pid>/backlog` with `{"text":"...","priority":"high|normal|low"}` — the field is `text`, not `title`, and a missing/blank one is a 400.

**Refer to items by their `key`, never by `id`.** Each item carries a
JIRA-style key: the project prefix, a dash, a zero-padded number — `MC-01`,
`MC-142`. The `id` is an 8-char uuid slice no human can read back, and a bare
`#12` is ambiguous as soon as two projects are on screen. Keys are never
recycled, so one always means the same item.

**Relations** (JIRA-style): `POST /api/project/<pid>/backlog/<item_id>/links`
with `{"type":"blocked_by","target":"MC-12"}`. Types: `blocked_by`,
`duplicate_of`, `continues`, `relates_to`. `target` takes `MC-12` (any case or
padding), `#12`, `12`, or a raw id — same project only; another project's
prefix 404s rather than matching the number. Post ONE direction; the inverse
(`blocks`, `duplicated by`, `continued by`) shows on the other item
automatically. Remove with `DELETE …/links?type=…&target=…`.

## Scheduler (Clayrune LOCAL — for jobs that outlive a session)

For long-term, repeatable jobs that should re-run an agent inside Clayrune after the current conversation ends.

- List: `GET /api/schedules`
- Create: `POST /api/schedules` with:
  ```json
  {"project_id":"...","task":"...","schedule_type":"daily|weekly|interval|once|cron",
   "time":"09:00","days":[],"interval_minutes":60,"run_at":"ISO8601","cron_expr":"..."}
  ```
- Update: `PUT /api/schedules/<id>`
- Delete: `DELETE /api/schedules/<id>`

**Picker rule:** if the job must still fire after this conversation ends → Clayrune scheduler. If it's a tight in-session poll loop (e.g. "check the build every 5 min") → Anthropic `/schedule` skill instead.

## Hivemind (multi-agent orchestration)

To create a hivemind (orchestrator decomposes a goal and spawns workers):

```bash
curl -s -X POST http://localhost:5199/api/hivemind/create \
  -H "Content-Type: application/json" \
  -d '{"project_id":"<pid>","goal":"GOAL TEXT","max_concurrent_workers":3,
       "orchestrator_model":"sonnet","worker_model":"sonnet"}'
```

Before creating, ask the user clarifying questions about scope, priorities, and constraints.

## Terminal pop-out

For visual commands, dashboards, or long-running processes the user should see:

```bash
curl -s -X POST http://localhost:5199/api/terminal/launch \
  -H "Content-Type: application/json" \
  -d '{"project_id":"<pid>","command":"<CMD>"}'
```

Output appears in Clayrune's terminal pop-out with full ANSI color.

## API discovery

When you need a Clayrune feature you haven't used before, do NOT guess endpoint names. Either:
- Grep `server.py` for `@app.route` to enumerate real endpoints, or
- `curl http://localhost:5199/` and inspect the served HTML.

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.