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 skills add cyanheads/uniprot-mcp-server --skill uniprot-guidegit clone --depth 1 https://github.com/cyanheads/uniprot-mcp-serverWrote 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/skills/cyanheads/uniprot-mcp-server/uniprot-guide)<a href="https://agentmods.dev/skills/cyanheads/uniprot-mcp-server/uniprot-guide"><img src="https://agentmods.dev/badge/skills/cyanheads/uniprot-mcp-server/uniprot-guide/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.
<a href="https://agentmods.dev/skills/cyanheads/uniprot-mcp-server/uniprot-guide"><img src="https://agentmods.dev/badge/skills/cyanheads/uniprot-mcp-server/uniprot-guide.svg" alt="Reviewed on agentmods" width="80" 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 | $0.00056 | $0.00967 |
| Opus 5.5 | $0.00022 | $0.00387 |
| Sonnet 5.5 | $0.00011 | $0.00193 |
| Haiku 4.5 | $0.00006 | $0.00097 |
Grade A, and why
uniprot-guide 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 4d 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 — 47 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Scope
Use UniProt for protein identity, functional annotations, sequence, cross-database identifiers, and reference-proteome context. The service is keyless. It does not provide sequence alignment, structure analysis, or a general UniParc retrieval tool. A cross-reference supplies an identifier for a downstream lookup; it does not supply that database's record.
Identifiers
| Identifier | Origin and use |
|---|---|
UniProtKB accession, such as P04637 |
Search results, UniProtKB-target mapping rows, or proteome proteins; identifies the entry to investigate. |
| NCBI taxon ID | uniprot_get_taxonomy or a protein's organism record; retain its name and rank when choosing the search scope. |
| Proteome UPID | uniprot_get_proteome; fixes the particular proteome for enumeration and comparison. |
| Isoform ID | An entry's isoforms or uniprot_get_sequence; keep it attached to isoform-specific conclusions rather than substituting the parent accession. |
| PubMed or database cross-reference | Entry annotation evidence or xrefs; preserve the database namespace when handing the identifier onward. |
Gates
- Establish identity before writing a dossier. Check the returned gene, organism, and reviewed status against the intended protein. A top-ranked name match or one-to-many identifier mapping is a candidate set, not permission to choose its first row. If candidates remain biologically distinct, narrow the query or ask which organism, strain, or protein the user means.
- Establish taxonomic scope before interpreting an empty result. Inspect the resolved name and rank. Exact organism nodes and their descendants are different search sets: species-level requests can need entries assigned to strains or subspecies. A reference proteome is a particular protein set, not every UniProt entry assigned to a species. Use the query reference to make the intended scope explicit.
- Check coverage before claiming absence. A reviewed-only search can miss unreviewed entries; a filtered proteome list can exclude the protein; a page can omit later matches. State the scope actually searched. A missing annotation means it was not returned or annotated in the retrieved record, not that the biological property is absent.
- Read evidence before making functional claims. Search function snippets omit evidence references. Fetch the entry annotations and retain their attached evidence for a cited summary. Distinguish reviewed status, annotation score, and protein-existence evidence instead of treating any one as proof of every functional claim.
- Finish the mapping set before reconciling it. Preserve source-to-target pairs and all candidate targets. A finished mapping job can still have further result pages; one page's mapping count is not a count of successfully mapped input IDs. Report unresolved inputs and any remaining pages before calling the translation complete.
What ships with it
1 file beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- 4d ago First seen · 47 lines · 56 tokens per session scan A a5fc83f6c41e
uniprot-guide is a skill published in the GitHub repository cyanheads/uniprot-mcp-server (2 stars, last pushed 5d ago), licensed Apache-2.0. It adds 56 tokens to every session and 967 once invoked, about $0.0002 per session on Opus 5.5. 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 2026-11-04.
uniprot-guide compared with skills from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| uniprot-guidecyanheads/uniprot-mcp-server | Skill | 2 | 5d ago | Apache-2.0 | A | 56 | Claude Code |
| protein-guidecyanheads/protein-mcp-server | Skill | 6 | 5d ago | Apache-2.0 | A | 95 | Claude Code |
| biomed-researchnickzren/biomed-agent | Skill | 16 | 1mo ago | Apache-2.0 | A | 39 | Claude Code, Codex |
| ensembl-guidecyanheads/ensembl-mcp-server | Skill | 3 | 10d ago | Apache-2.0 | A | 95 | Claude Code |
| gnomad-genetics-guidecyanheads/gnomad-genetics-mcp-server | Skill | 1 | 5d ago | Apache-2.0 | A | 53 | Claude Code |
| computasemadhusudan-kulkarni/computase | Skill | 0 | 2mo ago | MIT | A | 53 | Claude Code, Codex |
| noodle-biomedical-literature-discoveryhelena-bioinformatics/noodle-mcp | Skill | 0 | 2mo ago | Apache-2.0 | A | 105 | Codex |