What is SwarmAI context-and-memory-safety.md?

A set of instructions for loading an AI system's context and memory safely. It defines priorities for files such as user details, project information, stored memories, and domain knowledge, while allowing startup to continue if loading fails.

Use it when changing context-file loading, memory handling, prompt assembly, file copying, or token-budget rules.

It prevents missing or damaged context from stopping the agent and helps decide which information to keep when space is limited.

How xg-gh-25/SwarmAI configures Kiro

Reuse

Borrowing it

Nothing to install: this file belongs to xg-gh-25/SwarmAI. 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/xg-gh-25/SwarmAI/main/.kiro/steering/context-and-memory-safety.md

Clone the repo

git clone --depth 1 https://github.com/xg-gh-25/SwarmAI

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/xg-gh-25/swarmai/context-and-memory-safety/github.svg)](https://agentmods.dev/instructions/xg-gh-25/swarmai/context-and-memory-safety)

Your own site

<a href="https://agentmods.dev/instructions/xg-gh-25/swarmai/context-and-memory-safety"><img src="https://agentmods.dev/badge/instructions/xg-gh-25/swarmai/context-and-memory-safety/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/xg-gh-25/swarmai/context-and-memory-safety"><img src="https://agentmods.dev/badge/instructions/xg-gh-25/swarmai/context-and-memory-safety.svg" alt="Reviewed on agentmods" width="80" height="20"></a>

Per session 1,570 This file is loaded in full into every session.

When invoked 1,570 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	$0.01570	$0.01570
Opus 5	$0.00785	$0.00785
Sonnet 5	$0.00314	$0.00314
Haiku 4.5	$0.00157	$0.00157

Measured 4d ago against content hash 3f931a05d85e, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-10, from the pricing page.

Security

Grade A, and why

SwarmAI context-and-memory-safety.md 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 4d 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/context-and-memory-safety.md · 117 lines

How it starts

The opening of the file, as written. The whole thing — 117 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Context and Memory Safety Principles

Core Invariant: Context Loading Never Blocks Agent Startup

The entire context assembly pipeline (ContextDirectoryLoader.load_all() → _build_system_prompt() → SystemPromptBuilder.build()) is wrapped in try/except. A failure at any stage MUST degrade gracefully — the agent starts with reduced context rather than crashing.

The 11 Context Files — Priority and Ownership

P0  SWARMAI.md     System  Never truncated   App's core identity
P1  IDENTITY.md    System  Never truncated   Agent name/avatar
P2  SOUL.md        System  Never truncated   Personality/principles
P3  AGENT.md       System  Truncatable       Behavioral rules
P4  USER.md        User    Truncatable       User profile
P5  STEERING.md    User    Truncatable       Session-level rules
P6  TOOLS.md       User    Truncatable       Tool guidance
P7  MEMORY.md      User    Head-truncated    Cross-session memory
P8  EVOLUTION.md   User    Head-truncated    Self-evolution registry
P9  KNOWLEDGE.md   User    Truncatable       Domain knowledge
P10 PROJECTS.md    User    Lowest priority   Active projects

Two-Mode Copy Rules (ensure_directory)

These rules are the foundation of the user trust model. Violating them destroys user customizations.

System files (P0-P3, user_customized=False): ALWAYS overwrite on startup. Permissions set to 0o444 (readonly). These are the app's voice — updated on every release.
User files (P4-P10, user_customized=True): Copy ONLY if missing. Never overwrite. Permissions 0o644. These are the user's data.
Byte-comparison optimization: System files are only rewritten if content actually changed (avoids unnecessary disk IO and git noise).

Token Budget Invariants

Budget scales dynamically: ≥200K model → 40K tokens, 64K-200K → 25K, <64K → 25K (instance default)
Truncation order is strictly by priority DESC: P10 → P9 → P8 → ... → P3. P0-P2 are NEVER truncated.
MEMORY.md and EVOLUTION.md truncate from HEAD (discard oldest, keep newest at bottom)
All other truncatable files truncate from TAIL (keep beginning, discard end)
For models <32K, KNOWLEDGE.md and PROJECTS.md are excluded entirely (never enter truncation pipeline)

Read the full file on GitHub · 117 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.

4d ago First seen · 117 lines · 1,570 tokens per session scan A 3f931a05d85e

Subscribe to this mod's changes

SwarmAI context-and-memory-safety.md is an instructions file published in the GitHub repository xg-gh-25/SwarmAI (44 stars, last pushed 3d ago), licensed MIT. It adds 1,570 tokens to every session, about $0.0078 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 2026-09-06.

Building agents Memory and context agentic-ai ai-agent aws-bedrock claude claude-sdk context-engineering

next.js AGENTS.md

AGENTS.md instructions for vercel/next.js, covering next.js development guide, codebase structure, monorepo overview, core package: packages/next and other important packages.

vercel/next.js · 7,296 tokens

codex AGENTS.md

AGENTS.md instructions for openai/codex, covering rust/codex-rs, the codex-core crate, code review rules, crate api surface and model visible context.

openai/codex · 5,153 tokens

vscode buildNext.instructions.md

Working notes and architecture documentation for the new esbuild-based build system in build/next. Use when making changes to the new build pipeline (transpile/bundle commands, NLS plugin, source-map handling, resource copying, or self-hosting watch tasks).

microsoft/vscode · 6,785 tokens

SwarmAI: Instructions file for Kiro