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/berntpopp/metadome-link/agents-mdgit clone --depth 1 https://github.com/berntpopp/metadome-linkWrote 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/berntpopp/metadome-link/agents-md)<a href="https://agentmods.dev/instructions/berntpopp/metadome-link/agents-md"><img src="https://agentmods.dev/badge/instructions/berntpopp/metadome-link/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 | $0.02697 | $0.02697 |
| Opus 5 | $0.01349 | $0.01349 |
| Sonnet 5 | $0.00539 | $0.00539 |
| Haiku 4.5 | $0.00270 | $0.00270 |
Grade A, and why
metadome-link 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 — 190 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md — metadome-link
Guidance for agents and contributors working in this repository.
README.mdis the front door and follows the GeneFoundry README Standard v1 (hard ceiling 200 lines, fixed section order, machine-checked byscripts/check_readme.py). Depth lives indocs/— see the documentation index at the bottom of this file. Do not grow the README: relocate.
What this is
metadome-link is a data-reading MCP + REST server with an idempotent build trigger that wraps the MetaDome web
service (Wiel et al., Human Mutation 2019). It exposes, per human protein
transcript: the per-residue missense tolerance landscape (sw_dn_ds, lower =
more intolerant), Pfam domain annotations, meta-domain (homolog) variant
aggregation, and per-residue ClinVar annotations. MetaDome does not expose true
per-residue gnomAD counts; its explicitly-labelled Pfam aggregates may span genes
(GRCh37/hg19 or GRCh38.p14, defaulting to GRCh38.p14). It is one
backend in the GeneFoundry -link fleet, federated behind genefoundry-router
under the namespace metadome. It mirrors the sibling fleet stack lifted from
mondo-link (MCP plane) and mavedb-link (async API client + TTL cache).
Two planes (non-negotiable boundary)
- Data plane —
config.py,constants.py,identifiers.py,exceptions.py,api/,cache/,services/. Talks to MetaDome over async httpx, normalizes the responses, and returns plain dicts. It raises typed exceptions frommetadome_link.exceptions; it never builds error envelopes. Unlike most fleet siblings there is noingest/step and no local SQLite index of all data — MetaDome has no bulk dump, so this is a live-API proxy + persistent result cache (gtex/mavedb pattern). Completed landscapes are cached on disk (SQLite, keyed(transcript_id, metadome_data_version)); transcript lists use an in-memory TTL cache;/statusis never cached. - MCP plane —
mcp/. Domain-agnostic scaffolding shared with siblings (lifted frommondo-link).run_mcp_tool(inmcp/envelope.py) ownssuccess/_metaand converts exceptions into returned structured errors (never raised to the client).
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 Changed · +31 lines · +514 tokens per session 121e206cc90c
- 5d ago First seen · 159 lines · 2,183 tokens per session scan A 57c3b4b698bc
metadome-link AGENTS.md is an instructions file published in the GitHub repository berntpopp/metadome-link (0 stars, last pushed 3d ago), licensed MIT. It adds 2,697 tokens to every session, about $0.0135 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
claude-prompts-mcp CLAUDE.md
Claude Code instructions for minipuft/claude-prompts-mcp, covering claude prompts mcp -- operator handbook, core principles, node.js support boundaries, validation gates (one contract, impact-aware subsets) and fleet standards upstream (minipuft/repository-standards).
claude-prompts-mcp AGENTS.md
AGENTS.md instructions for minipuft/claude-prompts-mcp, covering agent rules · claude prompts mcp, canonical ownership and loading, core principles, node.js support boundaries and validation gates (one contract, impact-aware subsets).
pubcrawl CLAUDE.md
Claude Code instructions for nickjlamb/pubcrawl, covering claude.md, project overview, commands, testing and architecture.
gmail-mcp CLAUDE.md
Instructions for cunicopia-dev/gmail-mcp, covering gmail mcp — development guide, what it is, layout, module responsibilities and oauth model.
mcp AGENTS.md
Instructions for SpaceFrontiers/mcp, covering agents.md — space frontiers mcp server, what it does, layout, local dev and conventions.
mcpelevator CLAUDE.md
Claude Code instructions for pacnpal/mcpelevator, covering commands, architecture, conventions, agent skills and issue tracker.