Borrowing it
Nothing to install: this file belongs to mightbesaad/audit-event-mcp. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/mightbesaad/audit-event-mcp/main/CLAUDE.mdgit clone --depth 1 https://github.com/mightbesaad/audit-event-mcpWrote 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/mightbesaad/audit-event-mcp/claude-md)<a href="https://agentmods.dev/instructions/mightbesaad/audit-event-mcp/claude-md"><img src="https://agentmods.dev/badge/instructions/mightbesaad/audit-event-mcp/claude-md/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/instructions/mightbesaad/audit-event-mcp/claude-md"><img src="https://agentmods.dev/badge/instructions/mightbesaad/audit-event-mcp/claude-md.svg" alt="Reviewed on agentmods" width="80" 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 5.1 | $0.02188 | $0.02188 |
| Opus 5 | $0.01094 | $0.01094 |
| Sonnet 5 | $0.00438 | $0.00438 |
| Haiku 4.5 | $0.00219 | $0.00219 |
Grade A, and why
audit-event-mcp 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 10d 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 — 110 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CLAUDE.md — @kajaril/audit-event-mcp
This file is the constitution for Claude Code working in this repo. Read it before any non-trivial change.
What this repo is
@kajaril/audit-event-mcp is agent steward infrastructure: a hash-chained audit log backed by a Cloudflare Durable Object, with optional Merkle notarisation and a GDPR/AI-Act compliance skill. MIT-licensed.
This is a governance product. Every decision in the codebase is on the record. DPO buyers and regulators will read this code.
Required reading before structural changes
.context/kajaril-source-of-truth.md— strategy, stack, roadmap.context/kajaril-decisions-addendum-wk0.md— entity-level locked decisions.context/audit-event-mcp-build-plan.md— this repo's build plan + all locked decisions
These live in the workspace .context/ folder, not in this repo. Deviating from any locked decision requires stopping, asking, and recording the new decision before code changes.
Locked decisions
- Two Workers —
wrangler.jsonc(main) +wrangler.notary.jsonc(notary). They are independent deployments. Never merge them into one config. - DO SQLite, not D1 —
state.storage.sql.exec()is the API. Never add a D1 binding to this repo. new_sqlite_classes, notnew_classes— Wrangler migration tag for SQLite-enabled DOs. Do not change.- Shared R2 bucket — one
audit-payloadsbucket, per-client isolation via{client_id}/{event_id}key prefix indo.ts. Never per-client buckets. - CF Access JWT routing — Worker reads
payload.custom.client_idfrom the CF Access JWT; DO name =audit-do-{client_id}. Never use URL segments for client routing. - Hash chain algorithm — locked.
input_hash_slot = input_hash ?? input_hash_omitted_reason. Never empty string.chain_hash = SHA-256(id + "|" + event_type + "|" + input_hash_slot + "|" + (prev_hash ?? "")). Do not change without a new decision entry. - Six MCP tools exactly —
record_event,verify_chain,query_events,export_dossier,request_approval,check_approval. Was four; widened by the witness integration decision (kajaril-witness-integration-draft.md §3, Day 3, 2026-06-12). Do not add tools without a new decision. payload_refnever returned — tool responses never includepayload_refkeys or raw payload content. This is a privacy invariant.NOTARY_PRIVATE_KEYis a Workers Secret — set viawrangler secret put, never in source, comments, or commits.migrations/0001_init.sqlis applied bydo.ts— the DO constructor callsstate.storage.sql.exec(SCHEMA_SQL). Never apply viawrangler d1 migrations apply.- License + metadata — MIT.
package.jsonauthor, repository, homepage, bugs are locked. - EU jurisdiction pin (2026-07-10) — every tenant DO is created via
env.AUDIT_DO.jurisdiction("eu")before.idFromName(...), never the raw binding. All three call sites (src/index.ts,src/lib/approval-flow.ts,src/lib/audit-log.ts) must agree — a client whose ID is ever derived without the jurisdiction call would resolve to a different DO than one derived with it, silently splitting that tenant's chain in two. The R2 side needed no code change:AUDIT_PAYLOADSbinding now points ataudit-payloads-eu(created viawrangler r2 bucket create audit-payloads-eu --jurisdiction eu) instead of the original non-jurisdictionalaudit-payloadsbucket — binding name unchanged, so nothing indo.ts/main.tshad to change. Decided with zero pilot clients on the old bucket/namespace, so no data migration was needed; if that's no longer true when this is read, check first.
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.
- 10d ago First seen · 110 lines · 2,188 tokens per session scan A e1d1e38190b5
audit-event-mcp CLAUDE.md is an instructions file published in the GitHub repository mightbesaad/audit-event-mcp (2 stars, last pushed 2mo ago), licensed MIT. It adds 2,188 tokens to every session, about $0.0109 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
next.js AGENTS.md
AGENTS.md instructions for vercel/next.js, covering next.js development guide, codebase structure, monorepo overview, core package: packages/next and other important packages.
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.
vscode buildNext.instructions.md
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).
spec-kit AGENTS.md
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.
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).
langchain AGENTS.md
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.