Borrowing it
Nothing to install: this file belongs to heihei999/huasheng-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/heihei999/huasheng-mcp/main/CLAUDE.mdgit clone --depth 1 https://github.com/heihei999/huasheng-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/heihei999/huasheng-mcp/claude-md)<a href="https://agentmods.dev/instructions/heihei999/huasheng-mcp/claude-md"><img src="https://agentmods.dev/badge/instructions/heihei999/huasheng-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/heihei999/huasheng-mcp/claude-md"><img src="https://agentmods.dev/badge/instructions/heihei999/huasheng-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 | $0.01185 | $0.01185 |
| Opus 5 | $0.00593 | $0.00593 |
| Sonnet 5 | $0.00237 | $0.00237 |
| Haiku 4.5 | $0.00119 | $0.00119 |
Grade A, and why
huasheng-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 11d 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 — 89 lines — stays where its author put it; the contents beside it link to each section on GitHub.
xingce-solver Claude Code Rules
This project is a rule-based MCP / CLI solver for Huasheng Shisan 行测 methodology.
Do not use machine learning. Do not import sklearn, torch, tensorflow, statsmodels. Do not hardcode case IDs or entity-specific rules. Do not modify knowledge_base/all_cards.jsonl. Do not modify data_analysis.py. Do not modify CLI or MCP input schemas. Do not re-parse PDFs. Do not OCR. Do not call external LLM APIs. Do not use file diffs. Show full file content when reviewing. Commit only after tests pass.
For logic_reasoning:
- v5 argument reasoning must not regress.
- v6.1 translation reasoning must not regress.
- truth reasoning must be conservative.
- If truth assignments are not unique, return analysis_only.
- If multiple candidates tie, return analysis_only.
- Never choose the first candidate by list order.
- Never lower thresholds just to answer more cases.
For compose_xingce_answer_prompt (v0.4):
- MCP server does not call external LLM/API.
- Claude Code is the LLM executor.
- The tool only generates a strict answer prompt.
- Answer only when exactly one option is justified.
- Otherwise return analysis_only.
- Do not invent missing visual/table content.
- Do not guess. Do not default to A.
- Wrong = 0 has higher priority than more correct answers.
- v0.4 actual Claude Code MCP client regression passed.
- Actual visible MCP tools: 15.
- compose_xingce_answer_prompt is visible and works as intended.
- No answer / selected_option / prediction in tool output.
- Full regression report: outputs/actual_claude_code_mcp_v0_4_regression.md
For v0.4.1 conservative answer gate hardening:
- Person arrangement "左边/右边" now routes to logic_analysis (not graphic_reasoning).
- graphic_reasoning without image/visual_description blocks answer_allowed.
- data_analysis without material/table/material_text blocks answer_allowed.
- route_uncertain blocks answer_allowed.
- allow_answer=false always blocks answer_allowed.
- Blocked cases produce analysis_only.
- MCP still does not output final answer.
- No external LLM/API call.
- v0.4.1 actual Claude Code MCP client regression passed.
- Actual visible MCP tools: 15.
- compose_xingce_answer_prompt is visible and works as intended.
- No answer / selected_option / prediction in tool output.
- Full regression report: outputs/actual_claude_code_mcp_v0_4_1_regression.md
For v0.4.2 data material gate hardening:
- Material/table/chart signals ("表中", "根据表格", "上述资料") route to data_analysis.
- Independent material gate: material signals require material/table context.
- context_requirements.requires_table_or_material reflects material signal.
- No external LLM/API call. No solver/scaffold modification.
For v0.4.3 conservative route coverage hardening:
- Text-based arrangement questions (books, programs, contestants) route to logic_analysis.
- Definition questions with "所谓...是指...下列体现" pattern route to definition_judgement.
- Both require multiple signal types to prevent false positives.
- All v0.4.2 routing patterns and answer gates remain unchanged.
- No external LLM/API call. No solver/scaffold modification.
- Actual Claude Code MCP v0.4.3 regression passed. Actual visible MCP tools: 15.
- ChatGPT-side 60-case pressure test: 57/60 exact route matches, 60/60 safety gate passed.
- Top-level answer/selected_option/prediction leakage: 0/60. No safety-level bug found.
- Full regression report: outputs/actual_claude_code_mcp_v0_4_3_regression.md
- Pressure test report: outputs/v0_4_3_clean_candidate_60_case_pressure_eval_summary.md
For v0.5.0 module context override:
- Added module_hint / section_context parameters to route and compose tools.
- Supports Chinese section names (类比推理, 言语理解, 资料分析, etc.).
- module_hint overrides weak keyword routing; strong material signals still take priority.
- Returns v0.5.0 fields: module_hint, section_context, module_hint_applied, module_hint_conflict, heuristic_module_guess.
- Safety gates unchanged: missing_visual_content, missing_table_or_material, route_uncertain, allow_answer=false.
- MCP guidance tests: 245 passed (was 220, +25). Full pytest: 576 passed (was 551, +25).
- No external LLM/API call. No solver/scaffold/all_cards/cli modification.
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.
- 11d ago First seen · 89 lines · 1,185 tokens per session scan A 8ffee220efdd
huasheng-mcp CLAUDE.md is an instructions file published in the GitHub repository heihei999/huasheng-mcp (23 stars, last pushed 6d ago), licensed MIT. It adds 1,185 tokens to every session, about $0.0059 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 2026-08-30.
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.
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 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).