executing-plans
executing-plans
Use when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available
minecontextGuides MineContext/OpenContext runtime service, context workflows, configuration, and desktop packaging tasks.
Independent GitHub snapshot · Automated classification · Original documentation governs
Use this skill for the MineContext repository and its Python backend package,
whose installed distribution is MineContext, import package is opencontext,
and console entry point is opencontext.
MineContext is a local-first context-aware AI partner. Its backend captures screen, document, vault, and web-link context; processes that material into context types; stores it with SQLite plus ChromaDB or Qdrant; and exposes FastAPI routes, web pages, content-generation jobs, and chat/completion flows. The repo also contains an Electron/React desktop app that bundles the Python backend through PyInstaller.
Use this skill when the task asks to:
opencontext server or CLI;Do not use this skill for unrelated context-management libraries, generic
FastAPI guidance without MineContext-specific APIs, or ordinary Electron apps
that do not bundle the opencontext backend.
Confirm that the current task really names MineContext, OpenContext,
opencontext, or one of its repo-specific surfaces such as screenshot
capture, local context storage, generated todos/tips/reports, or the
Electron package.
If a checkout is present and freshness matters, read references/repo-provenance.md. Refresh this skill when the commit, dirty state, package version, or major public paths differ from that snapshot.
Verify the installed backend before trusting runtime advice:
python -m pip install -e .
python -c "import opencontext; print(opencontext.__version__)"
opencontext --help
The bundled scripts/check_runtime.py performs a broader import and CLI smoke check without starting the server.
If the task requires model inference, screenshot analysis, web-link capture, or generated insights, confirm whether the user has a working OpenAI-compatible or Doubao endpoint. Many workflows import offline but need credentials or a local model server for real generation.
| Task shape | Read next | Why |
|---|---|---|
Start the backend, inspect opencontext start, configure models/auth/storage, upload context, call API routes, or debug runtime behavior | runtime-service | Owns CLI, FastAPI, capture, processing, storage, context generation, and bundled runtime smokes. |
Build the backend executable, package the Electron app, diagnose pnpm, PyInstaller, helper binaries, signing, or platform packaging | desktop-packaging | Owns backend/desktop build flow and macOS helper binary guidance. |
| Understand system layers, data flow, and component boundaries before routing | references/architecture.md | Condensed architecture from the repo's own overview and source layout. |
| A failure crosses install/import/config/runtime/build boundaries | references/troubleshooting.md | Cross-cutting recovery before following a workflow-specific troubleshooting page. |
| Updating the managed repo-skill router entry after verification | references/repo-routing-metadata.json | Structured metadata consumed by the repo-skill importer. |
MineContext.opencontext.opencontext.opencontext.server.opencontext.OpenContext.opencontext start.1733 for the user-facing backend
debugger, while opencontext.cli falls back to configured web settings.${CONTEXT_PATH:.} and includes SQLite
document storage plus ChromaDB vector collections unless config switches to
Qdrant.frontend/package.json, electron-builder.yml, and
a prebuilt PyInstaller backend copied into frontend/backend/.CONTEXT_PATH values, local virtualenv paths, or user
data paths in generated artifacts or public notes.