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/0-soft/2captcha-mcp/claude-mdgit clone --depth 1 https://github.com/0-SOFT/2Captcha-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/0-soft/2captcha-mcp/claude-md)<a href="https://agentmods.dev/instructions/0-soft/2captcha-mcp/claude-md"><img src="https://agentmods.dev/badge/instructions/0-soft/2captcha-mcp/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 | $0.01260 | $0.01260 |
| Opus 5 | $0.00630 | $0.00630 |
| Sonnet 5 | $0.00252 | $0.00252 |
| Haiku 4.5 | $0.00126 | $0.00126 |
Grade A, and why
2Captcha-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 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.
How it starts
The opening of the file, as written. The whole thing — 108 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CLAUDE.md — Project rules for twocaptcha-mcp
This file is read on every Claude Code session opened in this repository.
What this project is
An MCP server that exposes the official 2Captcha API to Claude Code. 43 tools across:
- 31 captcha solvers (every method
AsyncTwoCaptchaexposes) - 5 management tools (
balance,report_good,report_bad,get_result,send_raw) - 3 pingback CRUD tools
- 3 webhook event-store tools (
pingback_events_list,pingback_events_get,pingback_events_clear) - 1 composite tool (
solve_and_wait_pingback)
Critical principles
- No mocks or workarounds in production code. Mocks live in
tests/only. - SOLID / KISS / DRY. Tool handlers should be ≤ 5 lines; every cross-captcha concern belongs in
schemas/common.py. - Root-cause fixes only. When a test fails, fix the underlying issue, never the symptom.
- First-principles thinking for every architectural decision.
Repository layout
twocaptcha_mcp/
├── __main__.py # stdio entrypoint
├── server.py # MCP Server bootstrap + dispatcher
├── config.py # pydantic-settings (.env)
├── logger.py # stderr-only (stdout is the MCP stdio channel!)
├── client/
│ ├── solver.py # SolverClient — wraps AsyncTwoCaptcha (DIP boundary)
│ ├── pingback.py # httpx → legacy res.php (no SDK helpers exist for this)
│ ├── rate_limiter.py # async sliding-window per-method rate limiter
│ └── errors.py # CaptchaApiError / Network / Timeout / Validation
├── schemas/ # one module per captcha family
├── tools/ # one module per captcha family + composite
└── webhook_receiver/ # Starlette app (twocaptcha-mcp-webhook console script)
tests/ # one test module per source module + parametrised solver tests
docs/
├── RELEASE.md # PyPI Trusted Publisher setup + per-release workflow
├── DEPLOY-WEBHOOK.md # production webhook deployment guide
└── release/ # repo metadata, awesome-mcp PR text, sample systemd unit, checklists
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 · 108 lines · 1,260 tokens per session scan A e0c8c4001341
2Captcha-MCP CLAUDE.md is an instructions file published in the GitHub repository 0-SOFT/2Captcha-MCP (3 stars, last pushed 1mo ago), licensed MIT. It adds 1,260 tokens to every session, about $0.0063 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
Agent-skills CLAUDE.md
Claude Code instructions for Captcha-Sonic/Agent-skills, covering captchasonic plugin for claude code, what this plugin does, mcp tools available, setup and skills.
mcp-context-forge AGENTS.md
AGENTS.md instructions for IBM/mcp-context-forge, covering agents.md, project overview, project structure, essential commands and setup.
api AGENTS.md
AGENTS.md instructions for vulnersCom/api, covering agents.md — using the vulners python sdk, install & authenticate, two ways to use it, 1. as a library and 2. as an mcp server.
yandex-tracker-mcp AGENTS.md
AGENTS.md instructions for aikts/yandex-tracker-mcp, covering agents.md, rules, project overview, commands and architecture.
python3-capsolver AGENTS.md
Instructions for AndreiDrang/python3-capsolver, covering agents.md, repository overview, where to work, architecture and boundaries and change rules.
cpf-validador CLAUDE.md
Claude Code instructions for opastorello/cpf-validador, covering claude.md, comandos, rodar localmente (fastapi + fastmcp na porta 8000), docker (porta 8002 conforme docker-compose.yaml) and arquitetura.