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/starfetch-js/starfetch/agents-mdgit clone --depth 1 https://github.com/starfetch-js/starfetchWhat 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 | $0.02436 | $0.02436 |
| Opus 5 | $0.01218 | $0.01218 |
| Sonnet 5 | $0.00487 | $0.00487 |
| Haiku 4.5 | $0.00244 | $0.00244 |
Grade A, and why
starfetch 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 yesterday.
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 — 263 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Starfetch Agent Guide
This file is contributor guidance for humans and coding agents working on Starfetch.
Project Boundary
Starfetch is a TypeScript/npm astronomical data querying project. The main surfaces are agent-ready TAP/ADQL access through a local MCP server, a scriptable CLI, and a distributable agent skill. The reusable TypeScript library powers those surfaces and is available for direct callers.
Keep the package boundaries tight:
@starfetch-js/coreowns TAP protocol behavior, VOSI metadata, VOTable parsing, output conversion, presets, registry discovery, async jobs, and typed errors.@starfetch-js/mcpis a thin MCP adapter over@starfetch-js/core.apps/mcp-appis a private Streamable HTTP host that constructs a fresh@starfetch-js/mcpserver and transport for each request. It owns only HTTP policy, request lifecycle, and process shutdown.@starfetch-js/skillowns packaged static skill files and install helpers.starfetchis the CLI package and should stay a thin caller over@starfetch-js/coreand@starfetch-js/skill.
Do not expand the current pre-1.0 scope into broad VO protocol support, Gaia-only workflows, GUI workflows, full ADQL parsing, or authenticated TAP flows unless a later issue explicitly changes scope.
The root README.md is the single human-facing repository guide. Package
READMEs are concise npm package-page summaries, and packaged skill Markdown is
a product asset consumed by agents and MCP resources. Do not create a parallel
docs/ hierarchy for user guides, demonstrations, or architecture notes unless
a later issue explicitly changes this policy. Keep private launch planning,
campaign tracking, raw media, and historical working notes outside the public
repository.
Runtime and Package Policy
The implementation is an npm workspace with TypeScript ESM packages:
- root private workspace package;
packages/coreis@starfetch-js/core;packages/cliisstarfetch;packages/mcpis@starfetch-js/mcp;packages/skillis@starfetch-js/skill.apps/mcp-appis the private@starfetch-js/mcp-appdevelopment host.
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.
- yesterday First seen · 263 lines · 2,436 tokens per session scan A da17db5802ac
starfetch AGENTS.md is an instructions file published in the GitHub repository starfetch-js/starfetch (0 stars, last pushed 1mo ago), licensed MIT. It adds 2,436 tokens to every session, about $0.0122 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
nina AGENTS.md
Instructions for isbeorn/nina, covering agents.md, scope, documentation boundary, repository knowledge and code style and formatting.
PyAutoArray AGENTS.md
Instructions for PyAutoLabs/PyAutoArray, covering pyautoarray — agent instructions, what this repo is, related repos, architecture and quick commands.
PyAutoArray copilot-instructions.md
Instructions for PyAutoLabs/PyAutoArray: You are working on PyAutoArray, the low-level data-structure and numerical layer of the PyAuto ecosystem (package autoarray).
PyAutoArray CLAUDE.md
Instructions for PyAutoLabs/PyAutoArray: The canonical, agent-agnostic instructions live in AGENTS.md. Claude Code loads them via the import below; if your tool does not process @-imports, open AGENTS.md in this directory and read it directly. @AGENTS.md.
telugu-calendar-utilities AGENTS.md
AGENTS.md instructions for socraticsurge/telugu-calendar-utilities, covering working agreement, frozen core — do not modify without explicit owner approval, how new features land and project rules.
nina CLAUDE.md
Instructions for isbeorn/nina, a project described as: N.I.N.A. (Nighttime Imaging 'N' Astronomy) - Advanced astrophotography sequencing software for Windows, supporting full automation of imaging sessions.