dsh-failure-capsule
Run the following command in DeepSeek Harness:
dsh plugin install YiHarvest/dsh-failure-capsule
Paste the following prompt into your AI chat to install this plugin:
In DeepSeek Harness, run 'dsh plugin install https://github.com/YiHarvest/dsh-failure-capsule' to install this plugin.
About this plugin
When using a Coding Agent in DeepSeek Harness, failures are often not isolated errors but involve a sequence of events like model requests, tool invocations, permission changes, and more. Traditional error logs might only capture the final crash point, missing crucial context leading up to the failure, which complicates diagnosis.
The dsh-failure-capsule plugin addresses this issue by acting as a local-first failure evidence package. It automatically captures pre-failure events, project state, runtime environment, and plugin combinations when a tool or agent fails. The plugin compiles these signals into a sanitized ZIP file containing a timeline, Git command outputs, plugin manifests, and more, with support for source map resolution to deobfuscate minified JS stack frames, giving developers a comprehensive view of the failure scenario.
This plugin is ideal for engineers and researchers using DeepSeek Harness locally. It operates entirely offline without modifying Harness core or uploading data, ensuring data security. By automating evidence collection, it helps quickly identify root causes, enhancing debugging efficiency while eliminating the hassle of manual evidence gathering.
Use Cases
- Automatically generating a failure evidence package when a tool call in DeepSeek Harness returns an error.
- Collecting pre-failure events and project state when an agent turn is blocked by policy or process.
- Used in local development environments to debug complex failure scenarios of Coding Agents.
Best For
- Software engineers using DeepSeek Harness for local development.
- Researchers who need to diagnose the causes of agent failures.
- Developers concerned with data privacy and local debugging efficiency.
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.