dsh-engram
Minimalist token-saving long-term memory plugin for DeepSeek Harness featuring zero-LLM capture and ESR evidence closure.
Install
$ - **安装方式已验证 / Verified install**: dsh plugin --profile web add dsh-engramPlugin 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
- Zero-LLM auto-capture of milestones, file edits, and error fixes using pattern matching with built-in secret redaction
- Symbolic index prompt injection ([ENGRAM] and [ESR] blocks) frozen per session to preserve KV-cache stability
- Deterministic BM25 in-process recall and session-start auto-recall over project workspaces
- ESR evidence-closure task tracking (draft -> active -> stable) with disk verification and relation graphing
- Context GC replacing default lossy LLM auto-compaction with mechanical eviction and pointer summaries
- Integrated DSH native settings and UI kanban board for memory and task visualization with zero external UI dependencies
Useful For
- Retaining project conventions, tool failure fixes, and structural decisions across multi-session DeepSeek Harness agent workflows
- Preventing agent hallucination of task completion through strict evidence requirements (artifacts, evaluations, memory refs)
- Reducing token consumption during prompt assembly and context compaction in long-running agent tasks
Who It Fits
- DeepSeek Harness developers seeking persistent memory across agent sessions
- Teams aiming to optimize token usage and KV-cache performance in multi-turn coding workflows
- Engineers requiring evidence-backed task progression and cross-session reproducibility
Documented Limitations
- Requires DeepSeek Harness runtime >=0.1.2-alpha.2 <0.2.0-0 and Node.js >=22.19.0
- Preset compaction auto-wiring requires running 'npx dsh-engram revert' prior to uninstalling to avoid dangling references
- Captures only tool result events and explicit engram_store invocations, intentionally ignoring conversation plumbing
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, 0 development, 7 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.
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
- skepsun/dsh-engram
- Registry source
- GitHub · dsh-plugin topic
- Registry classification
- Plugin
- Source checked
- 5f08d21 · 2026-10-22
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
- 6GitHub stars
- Forks
- 2
- Open issues
- 1
- Last commit
- 2026-10-19
- Last release
- 2026-10-19
Related DSH Plugins
Ranked by overlapping capabilities, use cases, plugin type, categories, and DSH profile.
Native SQLite-backed task management and project board bundle with Web UI and Agent workflow tools for DeepSeek Harness.
View plugin →Lingshu long-term auditable memory system and cognitive graph for DeepSeek Harness.
View plugin →Persistent multi-model agent teams with dynamic DAG orchestration, run monitoring, and token insights for DeepSeek Harness.
View plugin →Skill manager UI for DeepSeek Harness to discover, toggle, import, and update agent skills.
View plugin →