icm copilot-instructions.md

Instructions for using ICM, a persistent memory system for AI agents. They explain how to recall earlier context and when to store decisions, preferences, resolved errors, and completed work.

In plain words
What is it for?
Use them before starting a task to search memory and after important changes or discoveries to save a concise record.
Why use it?
They help the agent avoid repeating investigations and preserve useful knowledge between sessions.

Instructions file for GitHub Copilot

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 instructions/rtk-ai/icm/copilot-instructions
Clone the repo
git clone --depth 1 https://github.com/rtk-ai/icm

Made for: GitHub Copilot.

Per session 287 This file is loaded in full into every session.
When invoked 287 The same file — it is already loaded in full.
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 $0.00287 $0.00287
Opus 5 $0.00143 $0.00143
Sonnet 5 $0.00057 $0.00057
Haiku 4.5 $0.00029 $0.00029

Measured 2d ago against content hash 568be5c7b581, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

icm copilot-instructions.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 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.

.github/copilot-instructions.md · 33 lines

What it actually says

ICM — Persistent Memory for AI Agents

This project uses ICM for persistent memory across sessions.

Recall (before starting work)

Search for relevant past context before each task:

icm recall "query"                        # search memories
icm recall "query" -t "topic-name"        # filter by topic

Store (MANDATORY triggers)

You MUST call icm store when ANY of the following happens:

  1. Error resolvedicm store -t errors-resolved -c "description" -i high -k "keyword1,keyword2"
  2. Architecture/design decisionicm store -t decisions-{project} -c "description" -i high
  3. User preference discoveredicm store -t preferences -c "description" -i critical
  4. Significant task completedicm store -t context-{project} -c "summary" -i high

Do this BEFORE responding to the user. Not optional.

Other commands

icm topics                                # list all topics
icm health                                # topic hygiene audit
icm update <id> -c "updated content"      # edit memory in-place
icm forget <id>                           # delete a memory
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 · 33 lines · 287 tokens per session scan A 568be5c7b581

Subscribe to this mod's changes

icm copilot-instructions.md is an instructions file published in the GitHub repository rtk-ai/icm (549 stars, last pushed 3d ago), licensed Apache-2.0. It adds 287 tokens to every session, about $0.0014 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 instructions, from other repositories

ralph-loop-agent AGENTS.md

Instructions for vercel-labs/ralph-loop-agent, covering agent instructions, package management, ai gateway, source code reference and fetching additional source code.

vercel-labs/ralph-loop-agent · 532 tokens

vibe-garden CLAUDE.md

Instructions for rjroy/vibe-garden, covering claude.md, repository overview, repository structure, plugins and compass rose.

rjroy/vibe-garden · 705 tokens

vibe-garden copilot-instructions.md

Instructions for rjroy/vibe-garden, covering copilot code review instructions, confidence threshold, response format, do not comment on and language semantics you frequently misread.

rjroy/vibe-garden · 476 tokens

grit CLAUDE.md

Instructions for rtk-ai/grit, covering persistent memory (icm), recall (before starting work), store (after completing significant work), other commands and rtk (rust token killer) - token-optimized commands.

rtk-ai/grit · 1,479 tokens

azure-ml-finetuning-eval-skills copilot-instructions.md

Copilot instructions for kimtth/azure-ml-finetuning-eval-skills, covering github copilot custom instructions, available skills, training flow (azure-ml-llm-trainer), dataset generation (azure-ml-dataset-creator) and evaluation (azure-ml-model-evaluation).

kimtth/azure-ml-finetuning-eval-skills · 635 tokens

azure-ml-finetuning-eval-skills AGENTS.md

AGENTS.md instructions for kimtth/azure-ml-finetuning-eval-skills, a project described as: 🤖 Automate Azure ML finetuning with the LLM skill. 🔧.

kimtth/azure-ml-finetuning-eval-skills · 384 tokens