dsh-llm-fallbacks
Role-based automatic LLM retry and provider/model fallback strategy plugin for DeepSeek Harness.
Install
$ dsh plugin --profile web add dsh-llm-fallbacks # web profile (Settings → Fallbacks card)Plugin Overview & Capabilities
AI-assisted organization based on the public repository snapshot. The content must be grounded in source evidence and does not replace compatibility or security verification.
Key Capabilities
- Automatic failover across provider and model chains upon rate limits, auth errors, or exhausted retries
- Time-slot rotations to switch root chains based on peak and valley schedules
- Role-based fallback chains and multi-stage subagent role resolution
- Virtual 'FallbacksChain/Auto' provider in the model picker to use chains as the primary root route
- Configurable half-open recovery, cooldown durations, and per-step safety switch caps
- Full configuration support across Web GUI card, dsh-tui terminal commands, and settings.yaml
- Enforcement and reconciliation with DeepSeek Harness host subagent model selection policies
Useful For
- Preventing agent task failures during upstream provider outages, token exhaustion, or rate limiting
- Rotating model chains during peak vs. off-peak hours for cost optimization
- Assigning specific fallback sequences to subagents based on assigned persona roles
Who It Fits
- DeepSeek Harness agent developers requiring high model availability and fault tolerance
- Teams running multi-provider LLM setups with varying rate limits and peak-hour pricing
- Terminal and Web DSH operators managing automated subagent workflows
Documented Limitations
- Requires Node.js >= 22 and compatible DeepSeek Harness core versions
- Requires explicit 'enabled: true' configuration; defaults to inactive if omitted
- Tail of the all-day root chain must end in a conforming official V4 model (Flash or Pro)
DSH Compatibility
Version-specific runtime evidence collected by DSH Plugin. A missing result means we have not tested that combination yet.
Security Signals
Objective signals discovered from package metadata and source inspection. These are not a guarantee that a plugin is safe.
package.json declares 1 runtime, 33 development, 25 peer, and 0 optional dependencies.
package.json declares DSH bundle metadata.
GitHub reports the repository license as MIT.
A root package.json was captured and can be inspected by the registry.
package.json declares a prepare lifecycle script that may run during relevant package installation workflows.
Public GitHub source metadata is available for this registry snapshot.
Source & Registry Notes
Public provenance, Registry classification, and the latest source check for this entry, kept separate from runtime verification.
- Source repository
- omdsh-dev/dsh-llm-fallbacks
- Registry source
- GitHub · dsh-plugin topic
- Registry classification
- Plugin
- Source checked
- 6da889b · 2026-09-28
This project is independently indexed from public source information. DSH Plugin is not affiliated with DeepSeek or the plugin author. Always check the author repository before installation.
Repository Activity
- GitHub stars
- 20GitHub stars
- Forks
- 2
- Open issues
- 1
- Last commit
- 2026-09-27
- Last release
- 2026-09-27
Related DSH Plugins
Ranked by overlapping capabilities, use cases, plugin type, categories, and DSH profile.
Unofficial Command Code LLM provider plugin for DeepSeek Harness with live catalogs, account rotation, and web search.
View plugin →Connect ChatGPT (Codex), Claude, Grok, and GitHub Copilot subscriptions to DeepSeek Harness via OAuth.
View plugin →Persistent multi-model Agent teams and bounded DAG workflows with GUI controls for DeepSeek Harness.
View plugin →Multi-agent team collaboration plugin for DeepSeek Harness with DAG task scheduling, mailbox messaging, and an interactive Web UI monitor.
View plugin →