dsh-companion-dev
Run the following command in DeepSeek Harness:
dsh plugin install beijingwahw/dsh-companion-dev
Paste the following prompt into your AI chat to install this plugin:
Run dsh plugin install beijingwahw/dsh-companion-dev in your terminal to install; the source code is hosted at https://github.com/beijingwahw/dsh-companion-dev
About this plugin
Working daily on the DeepSeek Harness platform means juggling scattered conversation histories, opaque API spending, manual prompt iteration, and repeated context-switching for multi-model comparisons. dsh-companion-dev was built to consolidate these high-frequency, low-level tasks into a single plugin sandbox—no more hopping between five different tools.
Nine independently toggleable modules cover the full loop: smart conversation export (Markdown, PDF, JSON, and PNG long-image with built-in PII redaction and batch ZIP packaging); context handoff summaries (four-section generation injected into the next session as a system message); API cost optimization (a dynamic pricing engine that scrapes official rate pages hourly, peak/off-peak scheduling, smart model routing, and dual daily/monthly budget alerts with automatic non-essential call pausing); global and in-conversation search (fuzzy match, tag filtering, stream-safe highlight via the CSS Custom Highlight API); execution trace analysis (duration and token breakdown, anomaly flagging, and side-by-side diff reports); a Prompt engineering workbench (versioning, A/B testing, template library, and structured output JSON Schema validation); a multi-model arena (up to five models in parallel, batch leaderboards with P50/P95/P99 and cost scoring); breakpoint-resumable task orchestration (visual pipeline to YAML, progress persistence, cron and natural-language scheduling); and security and audit (multi-key management, DLP interception, operation logs, and compliance reports).
Every module is an independent Cordis sub-plugin—turn any combination on or off without side effects. All user data stays in the local Harness sandbox, API keys are encrypted at rest with AES-256-GCM, and the manifest explicitly disables telemetry. It is designed for solo developers who make DeepSeek Harness their primary workspace, small teams that need session-level collaboration and cost guardrails, and enterprise developers with compliance requirements—install one plugin and consolidate the most fragmented parts of your daily workflow.
Use Cases
- Batch-export multi-turn conversations with automatic PII redaction and ZIP packaging for archival instead of manual screenshots
- Automatically defer non-urgent API calls during official peak hours and receive daily/monthly budget alerts to avoid overspending
- Fire the same prompt at up to five models in parallel and compare output quality, latency, token usage, and cost side by side
Best For
- Solo developers who make DeepSeek Harness their primary workspace and frequently export or search conversations
- Small teams that need cross-session handoff, API cost guardrails, and iterative prompt engineering
- Enterprise developers with requirements for operation audit, DLP enforcement, and compliance reporting
Related Plugins
A method pack that makes AI coding agents plan against your real baseline, prove completion with fresh evidence, and reduce reworks and unsafe changes.
Turns the DeepSeek Harness session into a captain that builds a durable sub-agent team, splits goals into dependency-aware tasks, and coordinates work via direct messages and a live Web UI.
Gives coding agents design judgment, letting Claude Code, Cursor, and 70+ agents generate and iterate high-quality UI, presentations, and graphics right from the terminal.
Run the Pi ecosystem's plugins on DeepSeek Harness, unmodified, via a compatibility layer that implements Pi's public extension ABI on DSH's native services.