memory

A memory-management skill for storing and retrieving durable personal or project context across sessions. It supports recalling past information, explicitly saving new memories, reviewing historical sessions, and checking memory before commits.

In plain words
What is it for?
Use it to recall context, save user-requested memories, review past sessions, or evaluate project memory before a Git commit.
Why use it?
It prevents useful decisions and preferences from disappearing when a session ends while keeping frequently needed information separate from older notes.

Skill for Claude CodeCodex

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.

agentmods
npx agentmods add skills/wufunc/nova/memory
Any agent
npx skills add wufunc/nova --skill memory
Clone the repo
git clone --depth 1 https://github.com/wufunc/nova

Made for: Claude Code, Codex.

Per session 106 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 4,087 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
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.

ModelPer sessionOnce invoked
Fable 5 $0.00106 $0.04087
Opus 5 $0.00053 $0.02044
Sonnet 5 $0.00021 $0.00817
Haiku 4.5 $0.00011 $0.00409

Measured yesterday against content hash 7fd334980fca, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

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 yesterday.

The scan reads SKILL.md. This mod also ships 2 executable files (scripts/list_snapshot_candidates.py, scripts/search_notes.py), listed below but not scanned — reading those needs a real analyzer, not pattern matching.

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/memory/SKILL.md · 550 lines

How it starts

The opening of the file, as written. The whole thing — 550 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Memory Skill

Personal Memory Model

Personal memory under ~/.nova/memory/ is split into four distinct roles:

  • user.md — stable user identity and collaboration preferences
  • memory.md — curated "hot" long-term context that is worth loading first when a task needs memory
  • notes/ — atomic durable notes with frontmatter for grep-based retrieval
  • snapshot/ — episodic historical sessions, used only for explicit recall or later promotion

The key rule is:

memory.md is a curated hot set, not a second copy of notes/.


Mode: recall

Use this mode when the current task appears likely to benefit from prior durable memory. Typical triggers include non-trivial or multi-step work, multi-file changes, architecture or convention-sensitive tasks, work that depends on user or project history, recurring or ambiguous bugs, explicit memory requests, or any situation where you are unsure whether hidden past context matters.

Skip recall for simple self-contained tasks that can be completed safely from the user's message and immediate local context alone.

When this mode is used during session start, run it immediately after loading any relevant baseline memory files (user.md, project arch.md, and memory.md). If you initially skipped recall and later discover that more context is needed, run it before continuing. Use the user's request and current task framing as search input.

This mode searches durable memory only:

  • personal notes/
  • project ADR / DevLog

It intentionally does not search snapshot sessions during this durable path.

Step 1: Search Relevant Memories

Extract keywords from the task context (user request, technologies, file paths, error names, project names).

For the personal notes/ layer, default to a recent window of the last 7 days using note extracted_at and falling back to created_at.

Only widen the personal-note scope to all historical notes when the user explicitly asks for older, long-term, all-time, or historical memory.

Read the full file on GitHub · 550 lines

Files

What ships with it

2 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.

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.

  1. yesterday First seen · 550 lines · 106 tokens per session scan A 7fd334980fca

Subscribe to this mod's changes

memory is a skill published in the GitHub repository wufunc/nova (2 stars, last pushed 4mo ago), licensed MIT. It adds 106 tokens to every session and 4,087 once invoked, about $0.0005 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 2026-08-31.