What is evolving-memory?

A persistent memory system for scientific research sessions. It records findings, research ideas, strategies, and pitfalls so later sessions can reuse lessons from earlier work.

Use it to recall relevant strategies before an analysis and save verified findings or lessons after it finishes.

It reduces repeated research and helps avoid directions or methods that previously failed.

evolving-memory — Skill for Claude Code and Codex by Zaoqu-Liu

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

Any agent

npx skills add Zaoqu-Liu/ScienceClaw --skill evolving-memory

Clone the repo

git clone --depth 1 https://github.com/Zaoqu-Liu/ScienceClaw

Made for: Claude Code, Codex.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

README.md

[![agentmods](https://agentmods.dev/badge/skills/zaoqu-liu/scienceclaw/evolving-memory/github.svg)](https://agentmods.dev/skills/zaoqu-liu/scienceclaw/evolving-memory)

Your own site

<a href="https://agentmods.dev/skills/zaoqu-liu/scienceclaw/evolving-memory"><img src="https://agentmods.dev/badge/skills/zaoqu-liu/scienceclaw/evolving-memory/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

Your own site · 80×15

<a href="https://agentmods.dev/skills/zaoqu-liu/scienceclaw/evolving-memory"><img src="https://agentmods.dev/badge/skills/zaoqu-liu/scienceclaw/evolving-memory.svg" alt="Reviewed on agentmods" width="80" height="20"></a>

Per session 86 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.

When invoked 2,399 The whole file, excluding the scripts and references it only reads on demand.

Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.

Origin original No closer match found in the catalogue.

Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

Model	Per session	Once invoked
Fable 5.1	$0.00086	$0.02399
Opus 5	$0.00043	$0.01200
Sonnet 5	$0.00017	$0.00480
Haiku 4.5	$0.00009	$0.00240

Measured 5d ago against content hash 38b6f81175d8, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-08, from the pricing page.

Security

Grade A, and why

evolving-memory scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 5d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

skills/evolving-memory/SKILL.md · 224 lines

How it starts

The opening of the file, as written. The whole thing — 224 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Evolving Memory System

ScienceClaw learns from every research session. Each completed analysis enriches a persistent memory that makes future research faster, more accurate, and less error-prone.

Memory Record Types

All records are stored in ~/.scienceclaw/memory/findings.jsonl (append-only, one JSON object per line).

Type 1: Finding (existing)

A verified scientific discovery with evidence.

{"type":"finding","date":"2026-03-18","gene":"THBS2","disease":"pancreatic cancer","finding":"THBS2 overexpressed in 17/33 TCGA cancer types (Wilcoxon p<0.001)","significance":"high","sources":["PMID:32273438"],"tags":["expression","pan-cancer"],"project":"thbs2-tumor-2026-03-10"}

Type 2: Ideation

Records whether a research direction proved viable or not. Prevents revisiting dead ends and reinforces promising paths.

{"type":"ideation","date":"2026-03-18","direction":"THBS2+CA19-9 as pancreatic cancer diagnostic panel","feasibility":"low","reason":"Prospective AUC dropped from 0.96 to 0.69; biomarker validation failed in independent cohort","tags":["diagnostic","validation-failure","liquid-biopsy"],"project":"thbs2-tumor-2026-03-10"}

Fields:

direction: The research direction or hypothesis explored
feasibility: "high" | "medium" | "low" | "dead-end"
reason: Why this feasibility assessment was reached (with data)

Type 3: Strategy

Records an analysis approach that worked well (or better than the default). Enables progressively better methodology.

{"type":"strategy","date":"2026-03-18","task":"survival_analysis","strategy":"Use surv_cutpoint() for optimal expression cutoff instead of median split","outcome":"Better separation: p=0.003 vs p=0.047 with median; HR 2.31 vs 1.68","tools":["R:survival","R:survminer"],"tags":["survival","cutoff-optimization"],"project":"thbs2-tumor-2026-03-10"}

Fields:

task: The type of analysis (e.g., survival_analysis, enrichment, deseq2, literature_search)
strategy: What was done differently from the default approach
outcome: Quantitative evidence that this strategy was better
tools: Software/packages used

Read the full file on GitHub · 224 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

5d ago First seen · 224 lines · 86 tokens per session scan A 38b6f81175d8

Subscribe to this mod's changes

evolving-memory is a skill published in the GitHub repository Zaoqu-Liu/ScienceClaw (60 stars, last pushed 5mo ago), licensed MIT. It adds 86 tokens to every session and 2,399 once invoked, about $0.0004 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-03.

Memory and context Scientific

godot-optimization

Use when optimizing Godot games — profiler, draw calls, physics tuning, memory management, and common bottlenecks.

jame581/GodotPrompter · 28 tokens

evo-memory

Manages persistent research memory across ideation and experimentation cycles. Maintains two stores: Ideation Memory MI (feasible/unsuccessful directions) and Experimentation Memory ME (reusable strategies for data processing, model training, architecture, debugging). Three evolution mechanisms: IDE (after…

EvoScientist/EvoSkills · 186 tokens

swmm-modeling-memory

Read historical Agentic SWMM experiment audit artifacts and summarize repeated assumptions, QA issues, failures, missing evidence, run-to-run differences, lessons learned, and controlled skill update proposals. Use downstream of swmm-experiment-audit when multiple audited runs exist or when a user asks for modeling…

Zhonghao1995/agentic-swmm-workflow · 79 tokens

evolving-memory