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.
npx agentmods add instructions/threat-vector-security/guardian-agent/gemini-mdgit clone --depth 1 https://github.com/Threat-Vector-Security/guardian-agentWrote 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/instructions/threat-vector-security/guardian-agent/gemini-md)<a href="https://agentmods.dev/instructions/threat-vector-security/guardian-agent/gemini-md"><img src="https://agentmods.dev/badge/instructions/threat-vector-security/guardian-agent/gemini-md.svg" alt="Measured on agentmods" 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 | $0.02549 | $0.02549 |
| Opus 5 | $0.01274 | $0.01274 |
| Sonnet 5 | $0.00510 | $0.00510 |
| Haiku 4.5 | $0.00255 | $0.00255 |
Grade A, and why
guardian-agent GEMINI.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 5d 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.
This is a copy
86% identical to guardian-agent AGENTS.md — 9 lines differ, which has more behind it and is treated as the original. This page carries a canonical link to it rather than competing with it.
How it starts
The opening of the file, as written. The whole thing — 68 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Repository Guidelines
Branching (CRITICAL)
Do not create or switch to a new branch unless the user explicitly asks for it. Default to staying on the current branch, including main, unless the user directs otherwise.
Project Structure & Module Organization
src/ holds the TypeScript app; src/index.ts bootstraps the runtime. Subsystems live in src/runtime/, src/guardian/, src/tools/, src/channels/, src/llm/, and src/search/. Keep tests beside the code they cover as *.test.ts. The web UI is in web/public/. Use scripts/ for verification harnesses, docs/ for architecture and specs, policies/ for rule files, skills/ for bundled skills, and native/windows-helper/ for the Rust helper.
Build, Test, and Development Commands
npm run dev: start the app withtsx src/index.ts..\scripts\start-dev-windows.ps1 -StartOnly: start the Windows dev server without rebuilding for browser/API regression loops.bash scripts/start-dev-unix.sh --start-only: start the Unix/WSL dev server without rebuilding.npm run build: compile TypeScript todist/.npm run check: type check withtsc --noEmit.npm test: run the Vitest suite.npm run test:coverage: run Vitest with coverage thresholds.npx vitest run src/path/file.test.ts: run one test file.npm run validate:dependency-contract: validate pinned dependency and lockfile contract after dependency or override changes.npm run validate:windows-package: validate staged Windows package manifests after Windows packaging work.npm run helper:windows: rebuild the Windows native helper when touchingnative/windows-helper/.- Integration Testing (CRITICAL): Review
docs/guides/INTEGRATION-TEST-HARNESS.mdfor full-stack API regression testing, including guidance for cross-platform (Windows/WSL) and Ollama configurations.
Intent Gateway (CRITICAL)
All user intent classification must go through the Intent Gateway (src/runtime/intent-gateway.ts). Never use regex, keyword matching, string includes, or any ad-hoc pattern matching to determine what the user is asking for. The Intent Gateway is an LLM classifier that routes requests via structured tool calls. New routes must be added to IntentGatewayRoute, the tool schema, the system prompt, normalizeRoute, preferredCandidatesForDecision in direct-intent-routing.ts, and the candidate dispatch loop in src/index.ts. Pre-gateway interception is only permitted for slash-command parsing in channel adapters and continuation/approval flow detection.
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.
- 5d ago First seen · 68 lines · 2,549 tokens per session scan A c2f37bb92d5b
guardian-agent GEMINI.md is an instructions file published in the GitHub repository Threat-Vector-Security/guardian-agent (11 stars, last pushed 2mo ago), licensed Apache-2.0. It adds 2,549 tokens to every session, about $0.0127 per session on Opus 5. A static security scan graded it A with 0 findings. It is 86% identical to guardian-agent AGENTS.md, differing in 9 lines, and is treated as a copy.
Other instructions, from other repositories
agent-governance-toolkit copilot-instructions.md
Copilot instructions for microsoft/agent-governance-toolkit, covering copilot instructions for agent-governance-toolkit, architecture, build, test, and lint, python and install (editable, all extras).
continuum AGENTS.md
AGENTS.md instructions for shyftlabs/continuum, covering continuum — ai assistant knowledge pack, what is continuum?, repository layout, setup invariants (do not violate) and core api.
orloj AGENTS.md
Instructions for OrlojHQ/orloj, covering orloj agent instructions, must-follow sync rules and working style.
orloj CLAUDE.md
Instructions for OrlojHQ/orloj, covering orloj claude instructions, critical repository rules and implementation notes.
agent-starter-kit AGENTS.md
Instructions for ntorga/agent-starter-kit, covering agents.md, honesty and ambiguity, code shape, function extraction and duplication.
agentfootprint CLAUDE.md
Claude Code instructions for footprintjs/agentfootprint, covering agentfootprint — feature-work map, before you design it: it may already exist, module map, core state & flow and extension points.