context-graph-deepseek-harness
Run the following command in DeepSeek Harness:
dsh plugin install Mshir0/context-graph-deepseek-harness
Paste the following prompt into your AI chat to install this plugin:
Run dsh plugin install Mshir0/context-graph-deepseek-harness in your terminal to install; the full source code is at https://github.com/Mshir0/context-graph-deepseek-harness .
About this plugin
When you drive a coding agent on a real engineering codebase, the most wasteful thing you can do is either dump the entire repository into the prompt or let the model act on the current sentence alone. The first burns tokens; the second misses interface contracts and call-graph constraints.
The Context Graph plugin solves exactly that. It adds an interactive context-graph view inside DeepSeek Harness, turning functional goals, implementation files, and module dependencies into traceable nodes and edges. Core capabilities include incremental scanning of Python, C, and C++ modules, symbols, calls, and inheritance; targeted PDF chapter extraction guided by the document outline instead of loading the full text; and a pre-request context compilation step with token auditing that ensures only task-relevant context is injected, with a final boundary check on messages, snapshots, and tools.
Ideal for team or solo developers who use DeepSeek Harness for engineering-level work, maintain polyglot codebases with documentation assets, and want model context to be controlled, traceable, and auditable.
Use Cases
- Precisely inject task-relevant code context in polyglot projects
- Extract PDF documentation by outline section instead of loading full text
- Audit context and control token budgets before each model request
Best For
- Teams building polyglot projects with DeepSeek Harness
- Developers who want controllable, traceable, and auditable model context
- Developers maintaining medium-to-large projects with code and documentation assets
Related Plugins
Traceable, searchable cross-session memory for AI agents that turns conversation knowledge into a typed knowledge graph and recalls relevant subgraphs instead of replaying full history, natively integrated with DeepSeek Harness.
Gives DSH AI cross-session long-term memory, to-do and skill management, plus multi-session orchestration, external AI delegation, and an infinite canvas that grows with you.
dsh-mnemon is a three-tier, pluggable, Agent-driven memory system for DeepSeek Harness, combining Runtime memory, Project Documents, and replaceable Memory Spaces with nine long-term providers.
An opinionated, zero-infrastructure file-based memory protocol: human-readable markdown files plus a bounded always-loaded index, curation discipline, and no database or embeddings.