dsh-auto-memory
Persistent, file-based auto-memory plugin for DeepSeek Harness with prompt auto-injection and lifecycle eviction.
Install
$ dsh plugin --profile demo add dsh-auto-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
- Injects persistent memory index dynamically into the DeepSeek Harness system prompt under a configurable byte budget
- Provides agent memory tools including memory_write, memory_read, memory_list, memory_delete, memory_prune, and memory_delete_all
- Stores memories in human-readable, typed Markdown files with YAML frontmatter without external databases
- Supports pinned memories that resist budget truncation and staleness eviction
- Performs automatic background LLM consolidation of durable facts when a root session concludes
- Implements concurrency and crash protection with atomic cross-process file locks and orphan recovery
- Expands [[name]] cross-links during memory recall and evicts stale, unread memories automatically
Useful For
- Persisting user preferences, role details, and tech stack choices across distinct dsh sessions
- Retaining project-specific architectural decisions, debugging lessons, and workspace facts
- Automatically consolidating long chat session conclusions into reusable knowledge entries
- Managing and pruning accumulated agent memories safely without data corruption
Who It Fits
- DeepSeek Harness developers seeking persistent memory without running external vector databases
- Engineers requiring context continuity across multiple CLI/web sessions
- AI agent builders building specialized workflows with scoped project and user memories
Documented Limitations
- Requires DeepSeek Harness runtime environment (@deepseek-ai/dsh >= 0.1.5-rc.2) on Node ^22.19.0 or >=24.0.0
- Budget truncation under heavy memory load is positional rather than full semantic relevance-ranked
- External third-party vector/embedding search is not included; relies on file indexing and structured prompt injection
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 2 runtime, 14 development, 8 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
- AskTheWay/dsh-auto-memory
- Registry source
- GitHub · dsh-plugin topic
- Registry classification
- Plugin
- Source checked
- 5bcf1c6 · 2026-10-15
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
- 106GitHub stars
- Forks
- 7
- Open issues
- 3
- Last commit
- 2026-10-13
- Last release
- 2026-10-13
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 →Native SQLite-backed task management and project board bundle with Web UI and Agent workflow tools for DeepSeek Harness.
View plugin →Agent skill for DeepSeek Harness that turns ideas, plans, and codebases into interactive architecture diagrams.
View plugin →Second-model AI auto-review for DeepSeek Harness approval requests with fail-closed fallback and session audit logging.
View plugin →