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/rosselps/whyguard/agents-mdgit clone --depth 1 https://github.com/rosselps/whyguardWhat 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 | $0.04673 | $0.04673 |
| Opus 5 | $0.02337 | $0.02337 |
| Sonnet 5 | $0.00935 | $0.00935 |
| Haiku 4.5 | $0.00467 | $0.00467 |
Grade A, and why
whyguard 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 yesterday.
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.
How it starts
The opening of the file, as written. The whole thing — 302 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md
Development context for any coding agent working in this repository. Kiro and compatible agents load it automatically, alongside .kiro/steering/.
Where the detail lives:
| Topic | File |
|---|---|
| What the product is and is not | .kiro/steering/product.md |
| Components, pipeline, trust boundaries | docs/architecture/architecture.md |
| Evidence strength, confidence, block rule | .kiro/steering/evidence-policy.md |
| Package layout and dependency rules | .kiro/steering/structure.md |
| What the tool reads from a repository | docs/guides/feeding-whyguard.md |
| Coding standards and commit conventions | CONTRIBUTING.md |
Operating rules
- Read
.kiro/steering/before proposing architecture changes. - Preserve the modular-monolith, ports-and-adapters boundary (
domain→application→ adapters → apps). - Do not add a new service, database, framework, or integration without an ADR.
- Do not expand language support beyond TypeScript/JavaScript in the MVP.
- Do not build an IDE extension.
- Prefer deterministic Git/AST logic before LLM logic.
- Never claim a historical reason without evidence IDs.
- Distinguish protected behavior from current implementation.
- Never auto-execute generated tests.
- Keep GitHub permissions minimal — Contents/Pull requests/Issues/Metadata Read, Checks Read & Write only. Never request Contents Write for the App.
- Keep the CLI functional as a fallback — it remains a first-class surface alongside the MCP server.
- Include tests for every new detector or risk rule.
- Return
unknownwhen evidence is weak; never invent an incident, issue, or reason. - Do not weaken webhook, path, command, or secret-handling controls.
- Before finishing a task, run
pnpm typecheck,pnpm lint,pnpm format:check, and relevant tests. - The
.kiro/settings/mcp.jsonand.kiro/hooks/*.templatefiles in this repo are templates for a target workspace being protected by WhyGuard — never register a live PreToolUse hook against WhyGuard's own repository.
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.
- yesterday First seen · 302 lines · 4,673 tokens per session scan A 586e83a7b3da
whyguard AGENTS.md is an instructions file published in the GitHub repository rosselps/whyguard (0 stars, last pushed 1mo ago), licensed MIT. It adds 4,673 tokens to every session, about $0.0234 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 2026-08-31.
Other instructions, from other repositories
joern-mcp CLAUDE.md
Instructions for BlockSecCA/joern-mcp, covering joern-mcp, what this is, architecture, stack and commands.
CodeDebrief AGENTS.md
Instructions for ferdinandobons/CodeDebrief: This project uses CodeDebrief to keep decision flows synchronized with the source code.
CodeDebrief CLAUDE.md
Instructions for ferdinandobons/CodeDebrief: This project uses CodeDebrief to keep decision flows synchronized with the source code.
CodeDebrief GEMINI.md
Instructions for ferdinandobons/CodeDebrief: This project uses CodeDebrief to keep decision flows synchronized with the source code.
vscode design-philosophy.instructions.md
VS Code design philosophy — the shared Values→Principles→Moves vocabulary for reasoning about UI in design terms rather than raw pixels. Use when creating, editing, or reviewing any visual surface (CSS, DOM, theming, icons, motion). Name the value/principle before reaching for a token.
deepagents AGENTS.md
AGENTS.md instructions for langchain-ai/deepagents, covering global development guidelines for the deep agents monorepo, corridor security analysis, development workflow, suppressing ruff rules and pr conventions.