agentmods.dev

SCAM AGENTS.md

A project instruction file describing how AI coding agents should work on SCAM, a command-line benchmark for testing whether agents protect users from threats. It documents the project’s purpose, architecture, modules, and directory layout.

In plain words
What is it for?
Use it when modifying SCAM’s command-line interface, model adapters, simulated environment, evaluators, reports, exports, or related project files.
Why use it?
It gives an agent the project context and coding boundaries needed to make changes consistently. Without it, the agent may misunderstand how the benchmark and its simulated tools fit together.

Instructions file for CodexOpenCode

136 6mo ago A scan 3,027 tokens original MIT
View the source file 1Password/SCAM
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/1password/scam/agents-md
Clone the repo
git clone --depth 1 https://github.com/1Password/SCAM

Made for: Codex, OpenCode.

Per session 3,027 This file is loaded in full into every session.
When invoked 3,027 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 $0.03027 $0.03027
Opus 5 $0.01514 $0.01514
Sonnet 5 $0.00605 $0.00605
Haiku 4.5 $0.00303 $0.00303

Measured 2d ago against content hash f4759343acd4, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

SCAM AGENTS.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.

AGENTS.md · 270 lines

How it starts

The opening of the file, as written. The whole thing — 270 lines — stays where its author put it; the contents beside it link to each section on GitHub.

AGENTS.md — AI Coding Guidelines for SCAM

Project Overview

SCAM (Security Comprehension Awareness Measure) is a CLI benchmark that evaluates the safety of AI agents with tool access. It runs multi-turn conversations where an AI agent uses tools (inbox, browser, forms, credential vault) and must proactively protect the user from threats without being told to look for them.

Architecture

CLI (typer) → Agentic Runner → Model.chat() (tool-calling loop) → ToolRouter (simulated env)
                ↓                                                        ↓
         Checkpoint evaluator                                  Environment (emails, URLs, vault)
                ↓
         Reporting (rich + markdown)
                ↓
         Export (HTML, video, terminal replay)

Module Responsibilities

Module Purpose
scam/cli.py CLI commands: run, evaluate, replay, export, compare, report, scenarios
scam/models/base.py BaseModel (abstract: chat() required), ChatResponse, ToolCall
scam/models/__init__.py create_model() factory function, re-exports
scam/models/anthropic.py Anthropic Claude adapter (Messages API + tool calling)
scam/models/openai.py OpenAI adapter (Chat Completions API + tool calling)
scam/models/gemini.py Google Gemini adapter (google-genai SDK + function calling)
scam/models/discovery.py Dynamic model listing from provider APIs, interactive picker
scam/agentic/scenario.py YAML parser, AgenticScenario dataclass, STANDARD_TOOLS
scam/agentic/environment.py ToolRouter — simulates inbox, URLs, forms, vault; tracks dangerous calls
scam/agentic/evaluator.py Checkpoint-based scoring for agentic scenarios (regex + optional LLM judge)
scam/agentic/judge.py LLM-as-judge evaluator — semantic fallback when regex pattern matching misses
scam/agentic/runner.py Multi-turn conversation orchestrator for agentic evaluation
scam/agentic/reporting.py Terminal reports, comparison, markdown reports (agentic)
scam/agentic/aggregate.py Multi-run statistical aggregation (mean, std, CI, stability)
scam/agentic/replay.py Interactive terminal replay viewer with typing effects and scorecard
scam/agentic/export_html.py Self-contained HTML export with animated replay, per-scenario and combined index pages
scam/agentic/export_video.py MP4 video export using Pillow for frame rendering and FFmpeg for encoding. Includes title cards, animated message bubbles with markdown rendering, tool call visualization, and scorecard overlays
scam/utils/config.py Paths, model pricing, API keys, skill_hash(), agentic_scenario_hash(), estimate_agentic_cost(), calculate_cost()

Read the full file on GitHub · 270 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.

  1. 2d ago First seen · 270 lines · 3,027 tokens per session scan A f4759343acd4

Subscribe to this mod's changes

SCAM AGENTS.md is an instructions file published in the GitHub repository 1Password/SCAM (136 stars, last pushed 6mo ago), licensed MIT. It adds 3,027 tokens to every session, about $0.0151 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.

Data and AIPlanningTesting
Related

Other instructions, from other repositories

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 · 5,182 tokens

buildNext

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

next.js 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

vscode oss-third-party-notices.instructions.md

Instructions for microsoft/vscode, covering vs code oss third-party-notices pipeline, architecture, pipeline flow in ci, applying the notice (cutover) and fallback chain (never fail the build).

microsoft/vscode · 5,001 tokens

spec-kit AGENTS.md

Instructions for github/spec-kit, covering agents.md, about spec kit and specify, quickstart — add a new integration in 5 steps, integration architecture and integrationmanifest — file tracking.

github/spec-kit · 7,040 tokens

langchain AGENTS.md

Instructions for langchain-ai/langchain, covering global development guidelines for the langchain monorepo, corridor security analysis, project architecture and context, monorepo structure and development tools & commands.

langchain-ai/langchain · 4,345 tokens