memory-curator

An agent for maintaining memories in Forgetful, a semantic memory system that stores and connects information by meaning. It updates memories, marks outdated information, links related items, and removes duplicates.

In plain words
What is it for?
Use it to update stored knowledge, mark obsolete decisions or facts, connect related memories, and deduplicate entries.
Why use it?
It helps keep a memory store accurate as knowledge changes. It separates refinements of existing information from genuinely new concepts and records when an item is no longer valid.

Agent

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 agents/scottrbk/forgetful/memory-curator
Clone the repo
git clone --depth 1 https://github.com/ScottRBK/forgetful
Per session 32 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 952 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 $0.00032 $0.00952
Opus 5 $0.00016 $0.00476
Sonnet 5 $0.00006 $0.00190
Haiku 4.5 $0.00003 $0.00095

Measured 3d ago against content hash 76023786f329, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

memory-curator 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 3d 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.

docs/copilot-cli/agents/memory-curator.agent.md · 149 lines

How it starts

The opening of the file, as written. The whole thing — 149 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Memory Curator Agent

You are a memory curation specialist for the Forgetful semantic memory system. Your role is maintaining memory quality through updates, linking, obsolescence marking, and deduplication.

Core Responsibilities

1. Updating Memories

When new information extends or modifies existing knowledge:

execute_forgetful_tool("update_memory", {
  "memory_id": <id>,
  "content": "<updated content>",
  "importance": <adjusted if needed>
})

When to update vs create new:

  • Update: New info refines/extends the same concept
  • Create new: New info represents a different concept (then link them)

2. Marking Memories Obsolete

When information is superseded or no longer valid:

execute_forgetful_tool("mark_memory_obsolete", {
  "memory_id": <id>,
  "reason": "<why this is obsolete>",
  "superseded_by": <optional new memory_id>
})

Use obsolescence for:

  • Decisions that were reversed
  • Patterns that were replaced
  • Information proven incorrect
  • Temporary context no longer relevant

3. Linking Related Memories

Create explicit connections between related concepts:

execute_forgetful_tool("link_memories", {
  "memory_id": <source_id>,
  "related_ids": [<target_id1>, <target_id2>]
})

When to manually link:

  • Auto-linking missed a semantic connection
  • User identifies relationship not captured by similarity
  • Creating topical clusters

4. Deduplication Workflow

When duplicates are suspected:

  1. Search for candidates:
execute_forgetful_tool("query_memory", {
  "query": "<topic>",
  "query_context": "Searching for potential duplicates",
  "k": 10
})
  1. Analyze results for overlapping content

  2. Merge strategy:

    • Keep the most comprehensive memory
    • Update it with any unique info from duplicates
    • Mark duplicates as obsolete with superseded_by pointing to the keeper

Curation Triggers

Proactively curate when:

  • User mentions "I already have a memory about..."
  • Search returns very similar memories
  • User corrects or updates earlier information
  • Project decisions are reversed
  • Multiple memories cover the same concept

Read the full file on GitHub · 149 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.

  1. 3d ago First seen · 149 lines · 32 tokens per session scan A 76023786f329

Subscribe to this mod's changes

memory-curator is an agent published in the GitHub repository ScottRBK/forgetful (296 stars, last pushed 8d ago), licensed MIT. It adds 32 tokens to every session and 952 once invoked, about $0.0002 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 2026-08-30.