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/chawdamrunal/assay/claude-mdgit clone --depth 1 https://github.com/chawdamrunal/assayWrote 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/chawdamrunal/assay/claude-md)<a href="https://agentmods.dev/instructions/chawdamrunal/assay/claude-md"><img src="https://agentmods.dev/badge/instructions/chawdamrunal/assay/claude-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 | $0.02739 | $0.02739 |
| Opus 5 | $0.01370 | $0.01370 |
| Sonnet 5 | $0.00548 | $0.00548 |
| Haiku 4.5 | $0.00274 | $0.00274 |
Grade A, and why
assay CLAUDE.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.
How it starts
The opening of the file, as written. The whole thing — 107 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CLAUDE.md
This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.
What Assay is
Assay is a security scanner for the AI dev stack (Claude Code plugins, MCP servers, hooks, settings). It reasons about an artifact with an LLM ("LLM as judge, not janitor") rather than pattern-matching. It compiles to a single Go binary (bin/assay) with the React SPA embedded via embed.FS. The Go module is github.com/chawdamrunal/assay.
The default scan path drives the reasoning through the user's Claude Code subscription, not the Anthropic API — this is a hard product constraint (see README.md "How scans run"). The API path exists only as a --scan-mode legacy fallback.
Commands
Build, test, and lint go through the Makefile. Note it hardcodes GO := /opt/homebrew/opt/go/bin/go and expects golangci-lint and pnpm (via corepack enable) on PATH — golangci-lint is often not installed by default, so verify it locally before relying on make lint. Keep the Go-version pins in go.mod, the Makefile, and .github/workflows/ in sync before pushing.
make build # builds the SPA, copies it to internal/api/dist, then compiles bin/assay
make test # go test -race ./... AND cd web && pnpm lint (tsc --noEmit)
make lint # golangci-lint run ./... AND pnpm lint
make install # copies bin/assay to ~/.local/bin (override PREFIX=/usr/local)
make build runs make web first (pnpm install --frozen-lockfile && pnpm build), which is required because the Go binary embeds internal/api/dist/index.html — verify-embed fails the build if the embed copy didn't land.
Running a single Go test:
/opt/homebrew/opt/go/bin/go test -race ./internal/scanner/ -run TestName
/opt/homebrew/opt/go/bin/go test -race ./internal/verdict/ -run TestValidator/subtest
Three test suites, three cost profiles (build tags)
- Default —
go test ./.... All Sonnet calls useinternal/claude/fake.goFakeClient. Zero API cost. Gates every merge. - Integration —
go test -tags integration ./.... Replays recorded responses fromtestdata/recorded/against the golden corpus. Still zero API cost, full pipeline end-to-end. - Smoke —
go test -tags smoke ./.... Hits the real Anthropic API, needsANTHROPIC_API_KEY, ~$0.20/run. Excluded from CI; run manually before release. Lives ininternal/scanner/realapi_test.go.
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 · 107 lines · 2,739 tokens per session scan A 78cb9aacad3a
assay CLAUDE.md is an instructions file published in the GitHub repository chawdamrunal/assay (4 stars, last pushed 1mo ago), licensed Apache-2.0. It adds 2,739 tokens to every session, about $0.0137 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
SecureAI-Scan CLAUDE.md
Instructions for akanthed/SecureAI-Scan, covering claude.md, what this is, hard requirements (non-negotiable, apply to every change), 1. zero tolerance for false positives and commands.
aigis CLAUDE.md
Claude Code instructions for killertcell428/aigis, covering aigis — project guidelines, zenn記事管理（zenn cli + github連携）, ディレクトリ構成, 記事の作成 and プレビュー.
attackgen AGENTS.md
AGENTS.md instructions for mrwadams/attackgen, covering agents.md, project overview, development commands, running the application and installing dependencies.
ai-appsec AGENTS.md
Instructions for subodhkc/ai-appsec, covering agents.md — haiec agent security mcp (phase -1 → phase 18), 1. workspace layout, 2. architectural principles (apply to every phase), 3. hard constraints (never violate) and 4. phase discipline.
calllint copilot-instructions.md
Copilot instructions for calllint/calllint, covering github copilot instructions — calllint, invariants, quality gates (node 20 + pnpm), releases & copy and governance.
node9-proxy CLAUDE.md
Claude Code instructions for node9-ai/node9-proxy, covering node9 proxy — claude code rules, pr checklist (required before every commit), test rules, code rules and architecture notes.