summarize-skill
Run the following command in DeepSeek Harness:
dsh plugin install gtbwpkwjnb-alt/summarize-skill
Paste the following prompt into your AI chat to install this plugin:
Open a terminal or the plugin manager in DeepSeek Harness and run dsh plugin install gtbwpkwjnb-alt/summarize-skill to install this plugin; the source repository is available at https://github.com/gtbwpkwjnb-alt/summarize-skill .
About this plugin
Long-running coding agent sessions accumulate context that degrades performance, and cross-session handoffs still rely on manual copy-paste that drops critical decisions and constraints. summarize-skill splits the workflow into three distinct outputs: status check, context compaction, and task handoff, so an agent can produce a structured report with a single command at any point without re-reading the entire conversation.
The layout follows a conclusion-first, action-second, evidence-third principle. Issues are classified P0, P1, and P2. Every fact is tagged as observed, confirmed, inferred, estimated, or unavailable to keep speculation separate from verified information. The compact mode retains the minimal context needed to continue the current session; the handoff mode generates a paste-ready report for a new session including next steps and a reusable continuation prompt; the diagnostic mode adds evidence-based process risk observations on top of the handoff. By default everything is read-only: no files are created, no rules modified; only an explicit save command writes to the project directory and re-reads to verify.
Best suited for solo developers and small teams who run coding agents heavily: if you open three or four sessions a day, each requiring a cold-start recovery, or need to transfer a task from one agent to another, this skill cuts down repeated explanations and context waste significantly.
Use Cases
- Quick status check after a long agent session
- Cold-start recovery when resuming a task in a new session
- Handing off a task between different coding agents
Best For
- Solo developers who rely heavily on coding agents
- Small teams managing parallel multi-session workflows
- Tech leads focused on context-window efficiency and handoff quality
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.