Borrowing it
Nothing to install: this file belongs to whiteknightonhorse/APIbase. 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/whiteknightonhorse/APIbase/main/CLAUDE.mdgit clone --depth 1 https://github.com/whiteknightonhorse/APIbaseWrote 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/whiteknightonhorse/apibase/claude-md)<a href="https://agentmods.dev/instructions/whiteknightonhorse/apibase/claude-md"><img src="https://agentmods.dev/badge/instructions/whiteknightonhorse/apibase/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/whiteknightonhorse/apibase/claude-md"><img src="https://agentmods.dev/badge/instructions/whiteknightonhorse/apibase/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 | $0.00716 | $0.00716 |
| Opus 5.5 | $0.00286 | $0.00286 |
| Sonnet 5.5 | $0.00143 | $0.00143 |
| Haiku 4.5 | $0.00072 | $0.00072 |
Grade A, and why
APIbase 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 13d 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 — 30 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Language rule (ABSOLUTE) — ENGLISH-ONLY-1006, operator ruling 2026-10-05
English only, everywhere in this repository and in everything the product emits:
source code, comments, docstrings, UI text (HTML/Markdown/SVG), JSON/YAML config,
API/MCP responses, error messages, logs, Telegram notices, outgoing e-mail,
tests, docs, README, commit messages, branch and file names.
No Cyrillic characters anywhere in tracked files except the three Telegram slots below. No bilingual (RU/EN) copies.
Gate: tests/unit/english-only.test.ts — any Cyrillic in git diff means NOT DONE.
Outside the rule: the operator's own chat and ~/taskloop cards/rulings (outside this repo).
Telegram exception (ENGLISH-ONLY-1006 ruling-2, operator answer 2026-10-05: "TG in Russian")
The text of Telegram notifications the autopilot sends to the operator MAY be Russian. That text lives in exactly three slots and nowhere else:
- config/autopilot/tg-strings.ru.json -- string VALUES only (keys ASCII, flat object);
- config/autopilot/routing.json -- string elements of
variantsarrays only (they render as the Telegram "Variants:" line and, through the same function, in the operator file); - scripts/sync-counts-cron.sh -- the first argument of
calert "..."only. Everything else, in those files and in the whole repository, is English: code, comments, docstrings, JSON keys and_commentfields, notice()/notices.log lines, incident fields (what, system_did), operator files, fleet task files, tests, selftests. New Telegram text goes into tg-strings.ru.json (loaded at import, fail-closed), never inline. The gate tests/unit/english-only.test.ts enforces exactly these three rules; widening the allowlist needs a new ruling in ~/taskloop/disputes.
CodeQL rules (CODEQL-OPEN-ALERTS-1006 ruling-1, 2026-10-06): close alerts by code, never by dismiss
- Any function that inspects a request-derived value (.length, .slice, .startsWith, regex) first checks
typeof x === 'string'and declares the parameterunknown; Express query/body/header values can be arrays or objects (CodeQL js/type-confusion-through-parameter-tampering). - Every Express route whose handler reaches a DB/fs/auth call lists an express-rate-limit middleware literally in that route's handler array before the auth guard; a limiter inside a guard function or behind router.use is not counted and does not protect the pre-auth lookup (CodeQL js/missing-rate-limiting).
- Never strip multi-character markers (
<!--,<script,../) with a single String.replace; scan with indexOf in a loop until nothing changes, and still HTML-escape the result (CodeQL js/incomplete-multi-character-sanitization). Reference implementations: src/services/api-key.service.ts isValidApiKeyFormat (1), src/shop/routes/merchant.router.ts probeLimiter (2), src/shop/routes/legal.router.ts stripHtmlComments (3).gh api "repos/whiteknightonhorse/APIbase/code-scanning/alerts?state=open"must stay empty for these three rule ids.
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.
- 13d ago Changed · +6 lines · +288 tokens per session b4b719297733
- 19d ago First seen · 24 lines · 428 tokens per session scan A b2915146c9a3
APIbase CLAUDE.md is an instructions file published in the GitHub repository whiteknightonhorse/APIbase (10 stars, last pushed 2d ago), licensed MIT. It adds 716 tokens to every session, about $0.0029 per session on Opus 5.5. 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 2026-10-06.
Other instructions, from other repositories
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).
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.
vscode oss-third-party-notices.instructions.md
Instructions for microsoft/vscode, covering vs code oss third-party-notices pipeline, architecture, pipeline flow in ci, cache eligibility and rollout and applying the notice (cutover).
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.
vscode copilot-instructions.md
Copilot instructions for microsoft/vscode, covering vs code copilot instructions, project overview, root folders, core architecture (src/ folder) and built-in extensions (extensions/ folder).
deepseek-harness AGENTS.md
AGENTS.md instructions for deepseek-ai/deepseek-harness, covering agents.md, pre-stable apis and released session data, repository layout, commands and host sandbox failures.