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.
git clone --depth 1 https://github.com/MiguelMedeiros/unforgitWrote 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.
[](https://agentmods.dev/rules/miguelmedeiros/unforgit/unforgit-memory)<a href="https://agentmods.dev/rules/miguelmedeiros/unforgit/unforgit-memory"><img src="https://agentmods.dev/badge/rules/miguelmedeiros/unforgit/unforgit-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.
<a href="https://agentmods.dev/rules/miguelmedeiros/unforgit/unforgit-memory"><img src="https://agentmods.dev/badge/rules/miguelmedeiros/unforgit/unforgit-memory.svg" alt="Reviewed on agentmods" width="80" height="20"></a>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 | $0.00426 | $0.00426 |
| Opus 5.5 | $0.00170 | $0.00170 |
| Sonnet 5.5 | $0.00085 | $0.00085 |
| Haiku 4.5 | $0.00043 | $0.00043 |
Grade A, and why
unforgit-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 1mo 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
Unforgit Memory Integration
You have access to unforgit_recall and unforgit_add MCP tools for persistent repository memory across sessions.
1. Recall at the Start
At the beginning of every conversation, use the unforgit_recall tool with a query based on the user's first message to retrieve relevant context from past sessions.
Use recalled memories to avoid repeating past mistakes, follow known conventions, and apply previous decisions.
2. Save During the Conversation
Save memories as soon as something noteworthy happens — don't wait until the end.
Trigger a save with the unforgit_add tool immediately after:
- A bug is found and fixed
- An architectural or design decision is made
- A non-obvious gotcha or workaround is discovered
- A new convention or pattern is established
- A deployment/setup procedure is figured out
Memory types
| Type | When | Example |
|---|---|---|
semantic |
Decisions, conventions, architecture facts | "We use UTC timestamps everywhere" |
procedural |
Workflows, how-tos, playbooks | "To deploy: run make release, then kubectl apply" |
episodic |
Bugs found, gotchas, observations | "Found race condition in queue worker" |
Rules
- Keep text concise but self-contained — a future reader should understand it without extra context
- Use meaningful tags for discoverability (e.g.
["auth", "bug"],["deploy", "playbook"]) - Prefer
semanticfor stable facts,proceduralfor how-tos,episodicfor transient observations - Do NOT save trivial changes, obvious things, or info already in the codebase docs
- Quality over quantity — only save what's genuinely useful for future sessions
- Always write memory text in English, regardless of the language the user is speaking
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.
- 1mo ago First seen · 43 lines · 426 tokens per session scan A 42a224a90b12
unforgit-memory is a cursor rule published in the GitHub repository MiguelMedeiros/unforgit (11 stars, last pushed 2d ago), licensed MIT. It adds 426 tokens to every session, about $0.0017 per session on Opus 5.5. 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 cursor rules, from other repositories
atomicmemory
AtomicMemory persistent memory protocol and MCP tool usage.
brain
You have access to brain, a persistent memory layer shared across every session, every project, and every AI coding tool (Claude Code, Cursor, Codex, OpenClaw, Hermes). Notes written here survive session resets, context compaction, and tool switches. Git-backed, FTS5 search, 1ms query time.
cairn-commit-network
Commits that run cairn hooks need unsandboxed network.
memex
This project is connected to a persistent knowledge base (LLM Wiki) managed by the memex CLI.
agentic-stack
Portable brain via agentic-stack. Read .agent/ first.
ponytail
Ponytail, lazy senior dev mode. Always pick the simplest solution that works.