What is manna CLAUDE.md?

A set of project instructions for working on MANNA, an MCP server that exposes astronomical archive data to AI clients. The archives include services such as NOIRLab Astro Data Lab and NRAO/ALMA.

Use it when developing or testing the MANNA project. It covers dependency setup, tests, linting, server startup, Docker builds, MCP inspection, and environment settings.

It gives the coding agent project-specific commands, naming rules, architecture context, known pitfalls, and reliability requirements that general coding instructions would not provide.

How NSF-Simons-CosmicAI-Institute/manna configures Claude Code

Reuse

Borrowing it

Nothing to install: this file belongs to NSF-Simons-CosmicAI-Institute/manna. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.

Copy the file

curl -O https://raw.githubusercontent.com/NSF-Simons-CosmicAI-Institute/manna/main/CLAUDE.md

Clone the repo

git clone --depth 1 https://github.com/NSF-Simons-CosmicAI-Institute/manna

Made for: Claude Code.

Wrote 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.

README.md

[![agentmods](https://agentmods.dev/badge/instructions/nsf-simons-cosmicai-institute/manna/claude-md/github.svg)](https://agentmods.dev/instructions/nsf-simons-cosmicai-institute/manna/claude-md)

Your own site

<a href="https://agentmods.dev/instructions/nsf-simons-cosmicai-institute/manna/claude-md"><img src="https://agentmods.dev/badge/instructions/nsf-simons-cosmicai-institute/manna/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.

Your own site · 80×15

<a href="https://agentmods.dev/instructions/nsf-simons-cosmicai-institute/manna/claude-md"><img src="https://agentmods.dev/badge/instructions/nsf-simons-cosmicai-institute/manna/claude-md.svg" alt="Reviewed on agentmods" width="80" height="20"></a>

Per session 3,236 This file is loaded in full into every session.

When invoked 3,236 The same file — it is already loaded in full.

Security scan C 2 findings. A grade says what 26 rules found in the file — not that it is safe.

Origin original No closer match found in the catalogue.

Token cost

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	$0.03236	$0.03236
Opus 5	$0.01618	$0.01618
Sonnet 5	$0.00647	$0.00647
Haiku 4.5	$0.00324	$0.00324

Measured 11d ago against content hash 8dda1f7cc814, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-11, from the pricing page.

Security

Grade C, and why

manna CLAUDE.md scanned grade C with 2 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.

Cloud metadata endpointhighServer-side request forgery

One request to 169.254.169.254 can return temporary IAM credentials.

- **Every user-supplied URL clears `_url_guard.ensure_safe_url` before it is fetched.** `endpoint` (tap/cone/sia), `ivoid_or_url` (registry, when not an `ivo://` IVOID), and `job_url` (status/results/abort). The guard re

Makes network callslowCapability

Not a fault in itself. Listed so you know the mod talks to something, and to what.

- **`POST /mcp` 307-redirects to `/mcp/`** because of Starlette `Mount`. Inspector follows redirects; bare `curl /mcp` does not. Use `curl -L` or `/mcp/`.

CLAUDE.md · 126 lines

How it starts

The opening of the file, as written. The whole thing — 126 lines — stays where its author put it; the contents beside it link to each section on GitHub.

MANNA — Claude Code context

MANNA (MCP Architecture for NOIRLab and NRAO Archives) is an MCP server exposing IVOA-compliant astronomical archives (NOIRLab Astro Data Lab, NRAO/ALMA, …) to LLM clients. STABLE summer project (CosmicAI). Current version: 0.6.2 (stateless query-cache envelopes).

Casing: MANNA in prose; lowercase manna for every identifier — package, python -m manna, src/manna, image tags, and the MCP client alias. Never uppercase an identifier. Exception: the PyPI distribution is manna-mcp (bare manna is admin-prohibited on PyPI); the import package and CLI stay manna.

Commands

uv sync                                  # install deps + dev deps
uv run pytest --record-mode=none         # 682 tests, offline replay (incl. tests/evals/)
uv run pytest --record-mode=once -k <t>  # re-record one cassette (needs net)
uv run ruff check .                      # lint
uv run python -m manna                   # boot server on :8000 (MANNA_PORT to override)
docker build -t manna:dev .              # container build
npx -y @modelcontextprotocol/inspector --cli http://localhost:8000/mcp --method tools/list

Settings env vars are MANNA_* (Pydantic Settings, extra="ignore"). See .env.example.

Architecture

src/manna/
├── backends/          # TapClient, SiaClient, ConeClient, RegistryClient, ResolverClient
│                      # (typed pyvo/httpx/astropy wrappers — tools never import pyvo directly)
├── tools/
│   ├── tap.py                # vo_tap_query, vo_tap_status, vo_tap_results, vo_tap_abort
│   ├── archives.py           # vo_archive_list
│   ├── schema.py             # vo_schema_describe
│   ├── resolver.py           # vo_target_resolve
│   ├── registry.py           # vo_registry_search, vo_registry_describe
│   ├── cone.py               # vo_cone_search
│   ├── sia.py                # vo_sia_search
│   ├── find_observations.py  # vo_find_observations (cross-archive: resolve -> pick archive -> SIA/cone)
│   ├── count.py               # vo_count_observations
│   ├── survey.py              # vo_survey_target
│   ├── inspect.py             # vo_inspect_table
│   ├── _select.py             # shared resolve + archive-selection helpers for count/survey/find facades
│   └── _constants.py          # shared tool-layer constants (_ERROR_DOCSTRING)
├── archives/          # per-archive knowledge (one <short_name>.py each)
│   ├── _model.py      # Archive, Schema dataclasses (leaf)
│   ├── _select.py     # pure parse_allow/sort/select/validate helpers
│   ├── __init__.py    # registry: discover_archives() + get_active_archives()
│   ├── _endpoints.py  # endpoint lists/descriptions over the active set (Field examples)
│   ├── _knowledge.py  # per-table schema lookups (lookup_schema, active_schema_kb, schema_to_dict)
│   ├── _audit.py      # Audit dataclass: declarative live-probe spec read by evals/audit.py
│   ├── _traps.py      # trap push channels (silent cheatsheet + loud failure-time hint)
│   └── <archive>.py   # ARCHIVE = Archive(..., schemas=(...), priority=N)
│                      # (currently: alma.py, cadc.py, datalab.py, eso.py, gaia.py, gaia_ari.py, nrao.py, sdss.py)
├── _archive_label.py  # archive short_name lookup from a URL (label field on envelopes)
├── _url_guard.py      # SSRF guard: scheme + public-address check on every user-supplied URL
├── config.py          # Settings (MANNA_* env, extra="ignore") + get_settings() singleton
├── _serialization.py  # shared dataclass → JSON-friendly dict helper
├── shaper.py          # astropy.Table → inline envelope; oversize → result-URL/fetch_recipe
├── errors.py          # ToolExecutionError taxonomy + error_to_payload (spec §7)
├── observability.py   # JSON logging + current_request_id ContextVar
├── app.py             # build_mcp() + build_app() factories; RequestIdMiddleware
└── __main__.py        # uvicorn entry; called by `python -m manna`

Read the full file on GitHub · 126 lines

Changes

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 · 126 lines · 3,236 tokens per session scan C 8dda1f7cc814

Subscribe to this mod's changes

manna CLAUDE.md is an instructions file published in the GitHub repository NSF-Simons-CosmicAI-Institute/manna (2 stars, last pushed today), licensed MIT. It adds 3,236 tokens to every session, about $0.0162 per session on Opus 5. A static security scan graded it C with 2 findings (cloud metadata endpoint, makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.

Scientific

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.

vercel/next.js · 7,296 tokens

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.

openai/codex · 5,153 tokens

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).

microsoft/vscode · 6,785 tokens

manna: Instructions file for Claude Code