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/ccchenhuohuo/sorftime-mcp/agents-mdgit clone --depth 1 https://github.com/ccchenhuohuo/sorftime-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/ccchenhuohuo/sorftime-mcp/agents-md)<a href="https://agentmods.dev/instructions/ccchenhuohuo/sorftime-mcp/agents-md"><img src="https://agentmods.dev/badge/instructions/ccchenhuohuo/sorftime-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 | $0.01732 | $0.01732 |
| Opus 5 | $0.00866 | $0.00866 |
| Sonnet 5 | $0.00346 | $0.00346 |
| Haiku 4.5 | $0.00173 | $0.00173 |
Grade A, and why
sorftime-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 — 140 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Codex project instructions
This file is the authoritative development guide for coding agents working in this repository. The root README.md is written for repository users and operators; it is not a substitute for these instructions.
Mission
Maintain a governed Sorftime data service with one deterministic API core and three adapters:
- MCP is the team-facing execution, identity, audit, rate-limit, and policy boundary;
- the Sorftime Research Skill is the Host-side routing and interpretation protocol;
- the CLI is the administrator/developer surface for complete endpoint access, diagnostics, batch work, and emergency operations.
Probabilistic language understanding must end before endpoint authorization and execution begin.
Start here
Read the smallest relevant set before changing code:
- architecture or adapter boundaries:
docs/architecture.md; - MCP/Skill behavior:
docs/mcp-skill-integration.mdandskills/sorftime-research/SKILL.md; - deployment, authentication, secrets, or production readiness:
docs/deployment.md; - endpoint contracts:
src/endpoints.tsandsrc/core/governance.ts; - MCP tools:
src/mcp/server.ts; - CLI behavior:
README.mdandsrc/cli.ts.
Compatibility baseline
- package, MCP contract, and governance policy:
1.0.0; - endpoint registry: 52 Sorftime endpoints;
- ordinary MCP surface: 4 tools, free and read-only only;
- optional admin MCP surface: 2 tools, still free and read-only;
- Skill:
sorftime-research, compatible with MCP1.0.x; - runtime: Node.js 20+, TypeScript, MCP SDK v1, Streamable HTTP and stdio.
Non-negotiable invariants
- Never commit or print a real Sorftime Account-SK, MCP API key, proxy secret, Authorization header, or raw credential file.
src/core/governance.tsmust classify all 52 endpoints. MCP authorization never parses the humancoststring.- The initial MCP policy exposes only
billing=free,effect=read, explicitly allowlisted endpoints. - Paid reads, Coin-charged calls, task creation, updates, subscriptions, and deletes are not MCP tools.
- MCP never exposes
raw_callor arbitrary endpoint/body passthrough. Raw access remains CLI-only for operators. - User/tenant/role identity comes only from authenticated transport context. Tool inputs must never select or override identity.
- Stdio has no employee authentication; bind one principal at process launch and document it as local/operator mode.
- Audit records may contain actor, tenant, tool, endpoint names, marketplace, fingerprints, timing, decision, and outcome. They must not contain complete arguments, keywords, ASIN lists, headers, tokens, or upstream payloads.
- A free endpoint reporting positive
RequestConsumedopens the billing circuit. Do not continue automatically. - MCP retries remain zero. A lost response must not duplicate paid work; paid work is unavailable in MCP regardless.
- Preserve exact documented wire casing (
ASIN,Asin,Asins,Querystartdt). Do not normalize payload keys at the shared-client boundary. - Upstream unknown schemas remain JSON. Do not invent field meanings or present missing/unavailable values as zero.
- Shared quota is account-global, never an employee allowance. Per-person activity comes from MCP audit, not Sorftime balance fields.
- The Skill must not invoke the CLI, request credentials, invent identifiers, silently substitute stale monitoring for paid realtime data, or infer causality.
- CLI completeness must not weaken MCP policy. CLI and MCP share the API core, not the same public command surface.
- HTTP production/non-loopback mode must authenticate, restrict Hosts, validate Origins when present, cap sessions/concurrency, and keep the Account-SK server-side.
- MCP output and errors must remain sanitized. Unexpected upstream payloads never appear in public error details.
- Do not add experimental MCP Tasks for Sorftime async APIs until ownership, identity binding, TTL, persistence, and polling policy are designed and tested.
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 · 140 lines · 1,732 tokens per session scan A 1de5cc51584b
sorftime-mcp AGENTS.md is an instructions file published in the GitHub repository ccchenhuohuo/sorftime-mcp (1 stars, last pushed 1mo ago), licensed MIT. It adds 1,732 tokens to every session, about $0.0087 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
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.
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.
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).