What is memory-branching?

A workflow for making isolated branches of persistent memory, similar to branches in Git. It lets an agent test alternatives or risky changes without altering the main memory until the result is accepted.

Use it for technology evaluations, alternative approaches, or risky memory refactors that should be reviewed before becoming permanent.

It protects stable memory from experiments and makes it possible to compare, merge, or discard findings safely.

How matrixorigin/memoria configures Kiro

Reuse

Borrowing it

Nothing to install: this file belongs to matrixorigin/memoria. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.

Copy the file

curl -O https://raw.githubusercontent.com/matrixorigin/memoria/main/.kiro/steering/memory-branching-patterns.md

Clone the repo

git clone --depth 1 https://github.com/matrixorigin/memoria

Made for: Kiro.

Wrote 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.

README.md

[![agentmods](https://agentmods.dev/badge/instructions/matrixorigin/memoria/memory-branching-patterns/github.svg)](https://agentmods.dev/instructions/matrixorigin/memoria/memory-branching-patterns)

Your own site

<a href="https://agentmods.dev/instructions/matrixorigin/memoria/memory-branching-patterns"><img src="https://agentmods.dev/badge/instructions/matrixorigin/memoria/memory-branching-patterns/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.

Your own site · 80×15

<a href="https://agentmods.dev/instructions/matrixorigin/memoria/memory-branching-patterns"><img src="https://agentmods.dev/badge/instructions/matrixorigin/memoria/memory-branching-patterns.svg" alt="Reviewed on agentmods" width="80" height="20"></a>

Per session 992 This file is loaded in full into every session.

When invoked 992 The same file — it is already loaded in full.

Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.

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.

Model	Per session	Once invoked
Fable 5.1	$0.00992	$0.00992
Opus 5.5	$0.00397	$0.00397
Sonnet 5.5	$0.00198	$0.00198
Haiku 4.5	$0.00099	$0.00099

Measured 1mo ago against content hash d1f340c89004, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade A, and why

memory-branching 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.

.kiro/steering/memory-branching-patterns.md · 135 lines

How it starts

The opening of the file, as written. The whole thing — 135 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Memory Branching Patterns

Use Memoria's Git-like branching to isolate experiments, evaluate alternatives, and protect stable memory state.

Pattern 1: Tech Evaluation

Compare alternatives without polluting main memory.

memory_branch(name="eval_[technology]")
memory_checkout(name="eval_[technology]")

# Store findings on branch
memory_store(content="[technology] evaluation: [findings]", memory_type="semantic")

# When done — preview and decide
memory_diff(source="eval_[technology]")

# Accept: merge back
memory_checkout(name="main")
memory_merge(source="eval_[technology]", strategy="replace")
memory_branch_delete(name="eval_[technology]")

# Reject: just delete
memory_checkout(name="main")
memory_branch_delete(name="eval_[technology]")

Pattern 2: Pre-Refactor Safety Net

Snapshot + branch before risky memory changes.

memory_snapshot(name="pre_[task]", description="before [task]")
memory_branch(name="refactor_[task]")
memory_checkout(name="refactor_[task]")

# Do risky work on branch...

# If it goes wrong:
memory_checkout(name="main")
memory_branch_delete(name="refactor_[task]")
# main is untouched

# If it succeeds:
memory_diff(source="refactor_[task]")
memory_checkout(name="main")
memory_merge(source="refactor_[task]")  # default strategy: branch wins on conflicts
memory_branch_delete(name="refactor_[task]")

Pattern 3: A/B Memory Comparison

Two branches for competing approaches, diff both before deciding.

memory_branch(name="approach_a")
memory_branch(name="approach_b")

# Work on A
memory_checkout(name="approach_a")
memory_store(content="Approach A: [details]", memory_type="semantic")

# Work on B
memory_checkout(name="approach_b")
memory_store(content="Approach B: [details]", memory_type="semantic")

# Compare
memory_diff(source="approach_a")
memory_diff(source="approach_b")

# Merge winner, delete both
memory_checkout(name="main")
memory_merge(source="approach_a")  # default strategy: branch wins on conflicts
memory_branch_delete(name="approach_a")
memory_branch_delete(name="approach_b")

Read the full file on GitHub · 135 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.

1mo ago First seen · 135 lines · 992 tokens per session scan A d1f340c89004

Subscribe to this mod's changes

memory-branching is an instructions file published in the GitHub repository matrixorigin/memoria (623 stars, last pushed 2d ago), licensed Apache-2.0. It adds 992 tokens to every session, about $0.0040 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 2026-09-08.

Databases Git and workflow LLM Memory and context agent ai storage

memory-branching next to 6 instructions from other repositories
Mod	Type	Stars	Last push	Licence	Scan	Tokens	Works with
memory-branchingmatrixorigin/memoria	Instructions file	623	2d ago	Apache-2.0	A	992	Kiro
vscode buildNext.instructions.mdmicrosoft/vscode	Instructions file	195k	today	MIT	A	9,719	GitHub Copilot
next.js AGENTS.mdvercel/next.js	Instructions file	143k	today	MIT	A	7,461	Codex, OpenCode
langchain AGENTS.mdlangchain-ai/langchain	Instructions file	149k	today	MIT	A	4,680	Codex, OpenCode
vscode oss-third-party-notices.instructions.mdmicrosoft/vscode	Instructions file	195k	today	MIT	A	5,516	GitHub Copilot
vscode copilot-instructions.mdmicrosoft/vscode	Instructions file	195k	today	MIT	A	3,093	GitHub Copilot
deepseek-harness AGENTS.mddeepseek-ai/deepseek-harness	Instructions file	274k	2d ago	MIT	A	4,313	Codex, OpenCode

memoria: Instructions file for Kiro