memory-search

A command for searching saved project memories using both text matching and relationships in a knowledge graph.

In plain words
What is it for?
It is for searching Mnemosyne memories, with options to filter results or disable graph-based expansion.
Why use it?
It helps find relevant past information and optionally narrow results by namespace, importance, or result count.

Command for Claude Code

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 commands/rand/mnemosyne/memory-search
Clone the repo
git clone --depth 1 https://github.com/rand/mnemosyne

Made for: Claude Code.

Per session 13 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 643 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 $0.00013 $0.00643
Opus 5 $0.00006 $0.00321
Sonnet 5 $0.00003 $0.00129
Haiku 4.5 $0.00001 $0.00064

Measured 2d ago against content hash 73e12d8327f5, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

memory-search 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 2d 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.

.claude/commands/memory-search.md · 84 lines

What it actually says

I will search Mnemosyne memories and display formatted results.

Usage:

  • /memory-search <query> - Search with defaults
  • /memory-search --namespace <ns> <query> - Filter by namespace
  • /memory-search --min-importance <1-10> <query> - Filter by minimum importance
  • /memory-search --limit <N> <query> - Limit number of results
  • /memory-search --no-graph <query> - Disable graph expansion

Instructions for me:

  1. Parse the arguments:

    • Extract --namespace flag if present (otherwise auto-detect)
    • Extract --min-importance flag if present
    • Extract --limit flag if present (default: 10)
    • Extract --no-graph flag if present (default: expand_graph = true)
    • The remaining text is the search query
  2. Auto-detect namespace (if not specified):

    • Use Bash to check git root: git rev-parse --show-toplevel 2>/dev/null
    • Read CLAUDE.md for project name
    • Construct namespace as project:<name>
    • If no project: search all (null namespace)
  3. Call Mnemosyne MCP tool:

    {
      "name": "mnemosyne.recall",
      "arguments": {
        "query": "<search query>",
        "namespace": "<namespace or null>",
        "max_results": <limit>,
        "min_importance": <min_importance or null>,
        "expand_graph": <true/false>
      }
    }
    
  4. Format the output: For each result, display:

    Found <count> memories matching "<query>":
    
    <for each result, numbered>:
    
    <number>. [<importance stars>/10] <memory_type>
       Created: <date>
       Summary: <summary>
       Match: <match_reason> (score: <score>)
    
       <content preview (first 200 chars)>...
       Tags: <tags>
    
    

    Where:

    • Importance stars: ⭐ repeated times
    • Memory type: ArchitectureDecision, CodePattern, BugFix, etc.
    • Date: formatted as YYYY-MM-DD
    • Match reason: from the search result
    • Score: formatted to 2 decimal places

    If no results:

    No memories found matching "<query>"
    
    Suggestions:
    - Try different keywords
    - Remove namespace filter to search globally
    - Lower min-importance threshold
    - Check if memories exist: /memory-list
    
  5. Error handling:

    • If MCP server not available: "Error: Mnemosyne MCP server not running"
    • If query is empty: "Error: No search query provided. Usage: /memory-search "

Please search for memories using the provided arguments.

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. 2d ago First seen · 84 lines · 13 tokens per session scan A 73e12d8327f5

Subscribe to this mod's changes

memory-search is a command published in the GitHub repository rand/mnemosyne (84 stars, last pushed 9mo ago), licensed MIT. It adds 13 tokens to every session and 643 once invoked, about $0.0001 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.

Related

Other commands, from other repositories