Agents & sessionsSkill
EverOS
EverMind-AI
One portable memory layer for every AI agent: local-first, Markdown-native, user-owned, and self-evolving across apps, tools, and workflows.
Structure check pending
agent-memoryagentic-aiaichats
Back to Marketplace
Agents & sessionsPlugin
dsh-argp
yoza10635/dsh-argp
dsh-argp is a tool for DeepSeek Harness. Guarded context compaction for DeepSeek Harness (dsh): the LLM proposes, deterministic guards dispose — eager per-atom shrink (extract/summary/false under verbatim guards) + lazy reference-graph eviction (0-LLM) + byte-exact recall from an append-only log. Guarded context compaction: the LLM only proposes, deterministic guards decide — per-atom shrink + lazy reference-graph pruning (0-LLM) + byte-exact recall from an append-only log; compression ratio exactly delivered, history never destroyed.
- Stars
- 9
- Forks
- 1
- Issues
- 0
- Update
- 10 days
REVIEW SCORECARD
Review Scorecard
7.8/10Average
License compliance9.0
Engineering quality5.6
Content completeness7.0
Repository scale7.0
Maintenance activity9.0
✓ No risk detected · Automated scan results are for reference only, not an official endorsement
AI DEEP REVIEW
Worth AI Review
7.8/10Consider
A rigorous, well-tested 0-LLM compaction plugin with strong evidence, but niche and model-dependent.
Code quality8
README cites CI badge, `npm run check` (typecheck+smoke+unit) at 202/202 green, prefix-fingerprint and invariant tests, plus reproducible spike scripts with artifact paths.
Security7
No download-and-execute or plaintext secrets shown; install is via `dsh plugin add` and API keys come from env vars, but no security-specific documentation is provided.
Utility8
Targets a real pain point (lossy LLM-summarization compaction, cache invalidation, unbounded compression ratios) with measured 7/7 probe fidelity and exact budget enforcement, though value depends on model instruction-following.
Maintenance8
Last updated 2026-08-22, not archived, MIT licensed, with a CHANGELOG, releases, and dated test results (2026-09-02), indicating active development.
Documentation8
README is extensive with install commands, cordis.patch.yml config examples, architecture rationale, benchmark tables, reproduction commands, and an explicit known-limitations section, though it is dense and partly Chinese-first.
Generated by AI after reading the project README — a decision aid, not an endorsement. Neutral scores are given when information is limited.
PROJECT READMEREADMEExpandCollapse
Loading GitHub README...
The GitHub README is temporarily unavailable. Please try again later.
Open repositoryINSTALL REFERENCE
Install reference
dsh plugin --profile <name> add dsh-argpThis command was extracted from an explicit README install reference and is display-only; the DSH host will not execute it automatically.
PROJECT TOPICS
Project topics
VALIDATION LADDER
Validation progress
01Store discoveryPassed
02ClassificationPassed
03Structure checkPending
04Sandbox validationPending
FAQ
Frequently Asked Questions
What is dsh-argp used for?
Guarded context compaction for DeepSeek Harness (dsh): the LLM proposes, deterministic guards dispose — eager per-atom shrink (extract/summary/false under verbatim guards) + lazy reference-graph eviction (0-LLM) + byte-e
How popular is dsh-argp?
dsh-argp has 9 stars and 1 forks on GitHub, last updated 2026-10-11.
What license does dsh-argp use?
dsh-argp is distributed under the MIT license.
Is dsh-argp compatible with DeepSeek Harness?
dsh-argp is listed in the DSH Universe directory as a tool for DeepSeek Harness. This plugin is listed in the DSH Universe directory and covered by its validation pipeline.
CLASSIFICATION EVIDENCE
Classification evidence
Project typePlugin
Feature categoryAgents & sessions
Rule confidenceMedium
The system checks GitHub Topics against the site taxonomy and keyword rules. Current matches:context-compression。
Agents & sessionsSkill
MemOS
MemTensor
Self-evolving memory OS for LLM & AI Agents: ultra-persistent memory, hybrid-retrieval, and cross-task skill reuse, with 35.24% token savings and DeepSeek Harness support.
Structure check pending
agentagentic-aiaiai-agents
Agents & sessionsPlugin
harnessrouter
HarnessRouter
HarnessRouter Community Edition: the self-hosted, Apache-2.0 edition of the unified interface for agent harnesses. Run Codex, Claude Code, Hermes, PI, DSH, and more through one API, with sessions, streaming, files, cancellation, and failure handling. Implements the Unified Harness Protocol (UHP), an open standard. Your keys, your infrastructure.
Structure check pending
agent-apiagent-frameworkagent-harnessai-agents
Agents & sessionsSkill
memsearch
zilliztech
A persistent, unified memory layer for all your AI agents (e.g. Claude Code, Codex, DSH), backed by Markdown and Milvus.
Structure check pending
agentagent-memoryai-agentsclaude-code
Agents & sessionsPlugin
Tianshu-harness
huiliyi37
Tianshu is a terminal coding agent runtime built on the harness project (Tui X Gui), with prefix-cache engineering optimization for DeepSeek V4 (measured steady-state hit rate of 97–99% in long sessions) and deep adaptation. It breaks out of the traditional AI coding assistant's limit of treating the LLM merely as a tool, and is built on a cognitive virtual machine (CVM), a self-awareness layer and Stigmergy self-decaying memory, making the AI a development partner with independent judgment and cognitive safeguards.
Structure check pending
ai-agentcoding-agentcontext-managementdeepseek
Agents & sessionsSkill
modsearch
liustack
🥇 The strongest free web search plugin for DeepSeek Harness, and the search bridge for every model without native web access. Free, no signup, no API key. Ask the web or X, get structured JSON evidence. | 🥇 The strongest free web search plugin for DeepSeek Harness across the web, free with no signup and no API key. Adds search for models without web access; ask the web or X and get structured JSON evidence (search, scrape, cite).
Validation record available
agent-skillsagentic-workflowclaude-codeclaude-skills