"Run and debug the MineContext/OpenContext Python backend runtime,
Scanned 9/8/2026
Install to Claude Code
npx -y skills add VectorSpaceLab/AREX-Skill --skill runtime-service --agent claude-codeInstalls into .claude/skills of the current project.
Are you the author of Runtime Service?
Add the live security badge to your README — it updates automatically with every re-scan.
[](https://www.skillsdirectory.com/skills/vectorspacelab-runtime-service)More formats (shields.io, HTML) on the badges page.
---
name: runtime-service
description: "Run and debug the MineContext/OpenContext Python backend runtime,
APIs, configuration, storage, capture, processing, generation, chat, and local
smoke checks."
disable-model-invocation: true
metadata:
disco-role: operating
license: Apache 2.0
---
# Runtime service
Use this sub-skill for MineContext/OpenContext backend runtime tasks: starting
`opencontext`, configuring model/storage/auth/prompt settings, uploading or
capturing context, calling FastAPI endpoints, debugging content generation,
chat/completion flows, monitoring, and deterministic local smoke checks.
Route Electron app builds, PyInstaller packaging, signing, helper binaries, and
frontend release tasks to [desktop-packaging](../desktop-packaging/SKILL.md).
For cross-component orientation read the root
[architecture](../../references/architecture.md); for install/import or
runtime-vs-packaging failures read the root
[troubleshooting guide](../../references/troubleshooting.md) before the
workflow-specific troubleshooting page here.
## Runtime entry checklist
1. Confirm the installed backend package imports and the CLI is present:
```bash
python -c "import opencontext; print(opencontext.__version__)"
opencontext --help
```
2. Start from a known `CONTEXT_PATH` if you need disposable local data:
```bash
mkdir -p .minecontext-data
CONTEXT_PATH="$PWD/.minecontext-data" opencontext start --config config/config.yaml --host 127.0.0.1 --port 1733
```
3. Check health before calling data routes:
```bash
curl -s http://127.0.0.1:1733/health
curl -s http://127.0.0.1:1733/api/health
```
4. If API authentication is enabled, send `X-API-Key: <key>` or `?api_key=<key>`
on protected routes; never print or persist the key in notes.
5. Do not run screenshot, web-link, model-validation, generation, chat, or
completion workflows until the user confirms screen/browser/network/model
access and supplies non-secret credentials through config, environment, or the
settings API.
## Reference map
- [references/api-reference.md](references/api-reference.md): read when calling
HTTP routes, matching request bodies, interpreting response envelopes, or
debugging route-family ownership.
- [references/configuration.md](references/configuration.md): read before
editing YAML, `CONTEXT_PATH`, model settings, storage backends, prompt
language/files, content-generation intervals, or API auth.
- [references/data-formats.md](references/data-formats.md): read when creating
raw context payloads, interpreting processed contexts, selecting supported
document formats, or mapping storage tables and context types.
- [references/workflows.md](references/workflows.md): read for concrete CLI/API
workflows covering startup, uploads, screenshots, folder monitoring, web
links, context search, generation, chat/completions, monitoring, and local
smokes.
- [references/troubleshooting.md](references/troubleshooting.md): read when a
runtime workflow fails after the first health/config checks.
- [scripts/smoke_folder_monitor.py](scripts/smoke_folder_monitor.py): run for a
deterministic CPU-only folder-monitor create/update/delete check with mocked
storage cleanup.
- [scripts/smoke_document_text.py](scripts/smoke_document_text.py): run for a
CPU-only text-document processing smoke that builds an isolated temporary
runtime config and does not require model credentials.
## What this sub-skill owns
- `opencontext start`, `opencontext --help`, `OpenContext.initialize()`, capture
startup, FastAPI app state, and health checks.
- FastAPI route families for web pages, settings, context search, screenshots,
document/web-link uploads, vault documents, generated content, monitoring,
events, completions, context-agent chat, conversations, and messages.
- Runtime configuration: `config.yaml`, environment interpolation,
`CONTEXT_PATH`, model settings, prompt YAML, user settings, API auth, storage,
content-generation tasks, capture, and processing.
- Capture/processing/storage workflows for folders, documents, screenshots, web
links, ChromaDB, Qdrant, SQLite, and supported data formats.
- User-facing generation and consumption flows: activities, tips, todos,
reports, context-agent chat, streaming, message interruption, and intelligent
completion.
## Boundaries and safety
- Do not document or depend on the original repository checkout for normal use;
prefer installed `opencontext` plus bundled references/scripts.
- Do not include API keys, private model endpoints, local virtualenv paths,
package install prefixes, or user data paths in generated notes.
- Treat API settings updates, database resets, prompt imports, browser capture,
screenshot capture, and generation calls as user-data or credential-sensitive.
- Use [desktop-packaging](../desktop-packaging/SKILL.md) for executable/app
packaging tasks; this sub-skill only covers the backend after it is installed
or running.
Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.
No comments yet. Be the first to comment!