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.
npx agentmods add commands/tm42/mnemograph/remembergit clone --depth 1 https://github.com/tm42/mnemographWhat 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 | $0.00009 | $0.00490 |
| Opus 5 | $0.00005 | $0.00245 |
| Sonnet 5 | $0.00002 | $0.00098 |
| Haiku 4.5 | $0.00001 | $0.00049 |
Grade A, and why
remember 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.
What it actually says
Remember to Knowledge Graph
Store knowledge using the memory-store agent, which handles deduplication, canonical naming, and automatic relation creation.
Your Task
The user wants to remember: $ARGUMENTS
Instructions
-
Parse the user's input to identify:
- What they want to remember (the content)
- Any type hints (decision, pattern, learning, gotcha, question)
- Any mentioned relations to existing concepts
-
Spawn the memory-store agent with a structured request:
<store-request>
<item content="..." type_hint="..." related_to="..."/>
</store-request>
- Report the result to the user:
- What was stored (entity names, types)
- Any duplicates that were merged
- Any ambiguous matches the agent flagged
Type Detection Hints
| User says... | type_hint |
|---|---|
| "we decided", "chose", "decision" | decision |
| "pattern", "approach", "we use" | pattern |
| "gotcha", "learned", "turns out", "TIL" | learning |
| "question", "should we", "wondering" | question |
| "project", "codebase", "repo" | project |
| Generic fact | concept |
Examples
User: /remember we decided to use JWT for auth
Action: Spawn memory-store with:
<store-request>
<item content="we decided to use JWT for auth" type_hint="decision"/>
</store-request>
User: /remember gotcha: pytest needs python -m prefix
Action: Spawn memory-store with:
<store-request>
<item content="pytest needs python -m prefix" type_hint="learning"/>
</store-request>
User: /remember the API uses rate limiting, and we chose Redis for the cache
Action: Spawn memory-store with multiple items:
<store-request>
<item content="API uses rate limiting" type_hint="pattern"/>
<item content="chose Redis for cache" type_hint="decision" related_to="API"/>
</store-request>
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.
- 2d ago First seen · 71 lines · 9 tokens per session scan A e2a9ccd7286a
remember is a command published in the GitHub repository tm42/mnemograph (2 stars, last pushed 6mo ago), licensed MIT. It adds 9 tokens to every session and 490 once invoked, about $0.0000 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.
Other commands, from other repositories
ll
Extract lessons learned from the current session. Scans conversation for error→fix patterns, user corrections, and discoveries, then saves structured lessons to the wiki.
output
Generate output artifacts from active wiki content — summaries, reports, study guides, slide outlines, timelines, glossaries, comparisons. Outputs are filed back into the wiki.
al-memory-create
Generate or update memory.md file tracking decisions, changes, and learnings throughout project development for continuity across sessions. Use when you need to create or update memory, track decisions, or maintain session continuity.
advise
Pre-work command that queries past learnings and leverages MCPs before starting a new task.
retro
Post-work command that extracts learnings and updates the learning registry.
memorize.skeleton
Her sey kaydetmeye deger degildir. Hafiza, gelecekte TEKRAR KARSILASILDACAK bilgileri saklar. "Bunu bilseydim daha hizli yapardim" testini gec: ileride ayni durumla karsilasan ajan bu bilgiden faydalanir mi?