DeepSeek Harness PluginManifest Valid

metaflywheel

Persistent cognitive ledger and meta-problem modeling engine for DeepSeek Harness agent runtimes.

Developer ToolsSecurity & Policy

1GitHub Stars1ForksUpdated2026-10-18

Source

Install

No clear install command was found in the current repository evidence.

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.

source-grounded

metaflywheel is a resident cognitive flywheel engine designed for DeepSeek Harness (DSH). Instead of functioning like a generic vector memory store, it maintains a persistent, cross-session ledger tracking commitments (debts), completion criteria, lessons from past errors, and explicit capability boundaries. Entries are triggered and injected into the prompt dynamically when concrete runtime events occur, such as tool failures or invalid paths.

Key Capabilities

Tracks four core cognitive ledger categories across sessions: debts, completion criteria, failure lessons, and explicit boundaries
Triggers targeted memory recall based on concrete runtime error events rather than generic vector similarity
Provides a multi-tier self-verification system including audit tests and pre-commit consistency gates
Records audit trails and deposits directly to local workspace storage for transparency and inspectability

Useful For

Preventing repetitive mistakes across agent sessions by recalling prior corrections upon tool errors
Maintaining long-term accountability and unfinished task tracking across multiple conversational sessions
Enforcing strict completion definitions and behavioral constraints on autonomous agent workflows

Who It Fits

Developers building persistent agent workflows in DeepSeek Harness
Engineers requiring verifiable runtime error handling and long-term memory audits
Power users seeking consistent agent behavior that does not repeat known mistakes

Documented Limitations

Requires a DSH host runtime (mounted via Cordis patch) and cannot run as an independent background daemon
Does not run background autonomous processing when inactive; activates only upon user interactions and runtime triggers
Tracks user assertions and changes of mind as part of the shared ledger, which may introduce overhead for casual sessions
Does not modify underlying LLM weights or provide standalone vector similarity search

DSH Compatibility

Version-specific runtime evidence collected by DSH Plugin. A missing result means we have not tested that combination yet.

Not tested yetNo runtime compatibility tests have been published yet.

Security Signals

Objective signals discovered from package metadata and source inspection. These are not a guarantee that a plugin is safe.

Dependency Counts

package.json declares 0 runtime, 0 development, 2 peer, and 0 optional dependencies.

info

Dsh Bundle Declared

package.json declares DSH bundle metadata.

info

License Declared

GitHub reports the repository license as MIT.

info

Package Manifest Available

A root package.json was captured and can be inspected by the registry.

info

Source Available

Public GitHub source metadata is available for this registry snapshot.

info

Source & Registry Notes

Public provenance, Registry classification, and the latest source check for this entry, kept separate from runtime verification.

Source repository: NBagent-dev/metaflywheel
Registry source: GitHub · dsh-plugin topic
Registry classification: Plugin Bundle
Source checked: afe48d3 · 2026-10-18

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.