companion-memory
Native scoped Riko resident-memory and Wiki state-machine bundle for DeepSeek Harness.
Install
$ dsh plugin --profile web add /absolute/path/to/packages/bundle/riko-memoryPlugin 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
- Preset-scoped long-term memory isolation preventing cross-agent data leakage
- Multi-layered memory architecture (L0 Session Evidence, L1 Candidates, L2 Canonical Wiki, L3 Resident Snapshot)
- Explicit memory management tools: memory_remember, memory_correct, memory_forget, and memory_get_resident
- Background asynchronous Dream processing pipeline with failover support for Google Gemini, OpenAI, and Anthropic protocols
- Query-time hybrid recall with BM25 lexical ranking and optional dense vector retrieval
- Temporal validity tracking with historical snapshots, supersession lineage, and journaled session purge capabilities
- HTTP Control plane API and administrative inspection UI
Useful For
- Persisting user preferences and facts across multiple chat sessions for specific DSH agent presets
- Auditable enterprise agent memory with strict user-grounded confirmation rules
- Injecting compact, bounded resident memory context into LLM prompts without bloating token budgets
- Managing structured companion knowledge bases with conflict resolution and temporal validity
Who It Fits
- DeepSeek Harness agent developers building long-term conversational memory systems
- DSH workspace administrators requiring auditable and privacy-isolated memory storage
- AI companion creators needing structured resident state management
Documented Limitations
- Mandatory stable preset identity; missing preset configurations fail closed
- Tightly coupled to DeepSeek Harness workspace APIs and storageDomain rather than generic MCP/SQLite hosts
- Forget operations remove derived memory from future Residents but retain raw L0 session evidence unless explicit journaled purge is executed
- Live reranker is not wired to runtime Agent recall
- Public multi-node distributed consensus and background vector compaction are outside v1 scope
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 13 runtime, 11 development, 1 peer, and 0 optional dependencies.
package.json declares DSH bundle metadata.
GitHub did not report an SPDX license for this repository snapshot.
A root package.json was captured and can be inspected by the registry.
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
- AkinoHaruka/companion-memory
- Registry source
- GitHub · dsh-plugin topic
- Registry classification
- Plugin Bundle
- Source checked
- 86ac959 · 2026-09-27
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
- 0GitHub stars
- Forks
- 0
- Open issues
- 0
- Last commit
- 2026-09-24
- Last release
- No release detected
Related DSH Plugins
Ranked by overlapping capabilities, use cases, plugin type, categories, and DSH profile.
Knowledge graph memory and context engine for DeepSeek Harness with cross-session recall.
View plugin →Composable, view-based layered memory bundle for DeepSeek Harness with pluggable providers and strategies.
View plugin →LLM-supervised persistent memory and graph-based recall integration bundle for DeepSeek Harness.
View plugin →Lingshu long-term auditable memory system and cognitive graph for DeepSeek Harness.
View plugin →