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/rendyuwu/arknights-mcp/agents-mdgit clone --depth 1 https://github.com/rendyuwu/arknights-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/rendyuwu/arknights-mcp/agents-md)<a href="https://agentmods.dev/instructions/rendyuwu/arknights-mcp/agents-md"><img src="https://agentmods.dev/badge/instructions/rendyuwu/arknights-mcp/agents-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 5.1 | $0.01357 | $0.01357 |
| Opus 5 | $0.00678 | $0.00678 |
| Sonnet 5 | $0.00271 | $0.00271 |
| Haiku 4.5 | $0.00136 | $0.00136 |
Grade A, and why
arknights-mcp AGENTS.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 — 105 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md — Guardrails for coding agents
This file is the canonical, tool-agnostic instruction set for any AI coding
agent (Codex, Claude Code, and others) working in this repository.
CLAUDE.md points here.
Before changing files: read Arknights_MCP_PRD_v0.1_FINAL.md (the product
requirements) and SPEC.md (the distilled design of record). The
founder-approved decisions in PRD Section 18 (D1–D15) are binding; changing
one requires an ADR in docs/adr/ and explicit approval.
What this project is
A read-only Arknights intelligence MCP server: one shared application core,
two transports (local stdio, private OAuth/OIDC Streamable HTTP), over
versioned SQLite snapshots. Regions en and cn. Data is imported/synced via
CLI only; user-facing tools query SQLite and never touch the network.
Architecture rules (keep these separations)
- One shared core, two transports. Never duplicate domain logic across
transports — both call the same
tool_registry+ services (§V14). - Keep layers separated:
sources/(adapters),importers/field_policy.py(allowlist),importers/(parsers),db/+db/repositories/(data access),analyzers/(rules),models/+mcp/(schemas),transports/,auth/,middleware/. - Admin operations (
sync,import,validate,purge, source management) are CLI-only and are never exposed as MCP tools (§V28). - Write tests before broadening parsers. Prefer property tests (Hypothesis) for parsing/normalization.
Never (hard invariants — see SPEC.md §V)
- Make an outbound/source-network request from any MCP tool at query time (§V1); no query-time download/scrape fallback (§V24).
- Open SQLite writable in an MCP process; run arbitrary SQL, shell, filesystem, or a source-download tool; build SQL by string interpolation (§V2).
- Replace the active DB on a failed/incompatible sync, or mutate the active DB
in place. Promote only after validation, atomically via
current.json(§V3, §V4). - Emit a factual response without region + provenance, or silently mix
enandcn(§V5). - Emit an observation without
rule_id+ evidence + confidence + limitations +analyzer_version(§V6). Match on typed fields, not NL keywords (§V26). - Label a recommendation "mandatory" or "universal best"; recommend anything at
confidence
< 0.5(report it as a limitation instead) (§V7, §V8). - Concatenate imported/source text into server instructions or tool descriptions; treat imported strings as anything but untrusted data; generate tool descriptions from imported content; let source data select tools, URLs, or admin actions. Strip control characters and cap lengths (§V18, PRD 17.6).
- Log full prompts, full tool arguments, response bodies, authorization headers, bearer tokens, raw source records, or roster/account data; leak stack traces or local paths in errors (§V12, §V23).
- Request, store, or transmit Arknights game credentials or account IDs (§V15).
- Commit or ship raw snapshots, prebuilt databases, artwork, audio, story
scripts, voice lines, wiki/community prose, or full announcement bodies
(§V16).
data/builds/and*.sqliteare git-ignored. - Run non-loopback remote without HTTPS + valid OAuth/OIDC; implement username/password storage; enable public access via a single flag/env var (§V9, §V10; PRD 17.7).
- Provide bulk-dump / DB-download / unbounded pagination / enumeration; exceed search default 10 / max 50, page_size max 100, or the 200 KB default response cap (§V19, §V22).
- Upgrade the MCP SDK to v2, break a required tool field, or bump
schema_versionwithout an ADR (§V21, §V25). - Assume reuse permission from a public repo, an attribution offer, or a takedown offer (D13).
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 · 105 lines · 1,357 tokens per session scan A f610dc50cf20
arknights-mcp AGENTS.md is an instructions file published in the GitHub repository rendyuwu/arknights-mcp (0 stars, last pushed 1mo ago), licensed Apache-2.0. It adds 1,357 tokens to every session, about $0.0068 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
unity-code-style-guide AGENTS.md
Instructions for krogh-jacobsen/unity-code-style-guide, covering agents.md — unity 6 c, project setup — edit this block, never do these — they corrupt the project, deprecated in unity 6 and if you read nothing else.
nes-php-glfw copilot-instructions.md
Instructions for oliverearl/nes-php-glfw, covering github copilot development guidelines for nes emulator, agent document source of truth, project overview, code style & standards and documentation.
game-and-watch-retro-go-sd CLAUDE.md
Claude Code instructions for sylverb/game-and-watch-retro-go-sd, covering claude.md, what this project is, build / flash workflow, architecture and three storage tiers, one elf.
base-building-trap-defense-design-agent-skill CLAUDE.md
Instructions for dungnotnull/base-building-trap-defense-design-agent-skill, covering claude.md — skill 250: base-building-trap-defense-design, skill identity, problem this skill solves, harness flow summary and sub-skills.
trident-mcp AGENTS.md
Instructions for mordor-forge/trident-mcp, covering agents.md, what this is, build and test, single-file verification and e2e tests.
unreal-mcp-additional-tools CLAUDE.md
Claude Code instructions for nodormu/unreal-mcp-additional-tools, covering claude.md, what this is, commands, architecture and transports (src/transports/).