v0.2.1: entry-resource outline + identifier/notice fixes

entry-resource outline + identifier/notice fixes

A resource-overflow safeguard plus two protein-tool correctness fixes.

Changed:

• uniprot://entry/{accession} outlines a record over the budget — bounded identity summary + section outline (names + byte sizes) that routes to uniprot_get_entry (sections:[...]) instead of injecting the full payload (#6)

Fixed:

• uniprot_get_proteome and uniprot_get_taxonomy reject conflicting identifiers with a typed conflicting_identifier error (#5)
• uniprot_get_proteome declares the truncation notice field so paging guidance reaches content[] for text-only clients (#4)

113 tests pass; bun run devcheck clean.

v0.2.0: Flatten uniprot_get_entry output (breaking)

Flatten uniprot_get_entry output (breaking)

Breaking relative to 0.1.1: uniprot_get_entry no longer wraps its output in a result key — read structuredContent.succeeded[]/failed[]/sections[] directly.

Changed:

• uniprot_get_entry output flattened (breaking) — result wrapper and discriminated union removed; output is a flat object with top-level kind ("full" | "outline") plus optional succeeded[]/failed[]/sections[]. Outline re-call guidance moves to enrichment.notice. (#2)

Fixed:

• uniprot_get_sequence isoforms — include_isoforms now routes through the search endpoint, which honors includeIsoform; the per-accession .fasta endpoint silently ignored the flag and returned only the canonical record. (#1)
• uniprot_search_proteins custom fields — mandatory identity/provenance columns (MANDATORY_SEARCH_FIELDS) merged into every search request so a trimmed projection can no longer crash output validation on the missing entryName. (#3)

105 tests pass; bun run devcheck clean.

v0.1.1: Initial release — protein-first research over UniProtKB

Initial release — protein-first research over UniProtKB

Keyless protein research over the UniProt REST API — UniProtKB search and curated entries, cross-database ID mapping, reference proteomes, taxonomy, and sequences.

Added:

• 6 tools: uniprot_search_proteins, uniprot_get_entry, uniprot_map_ids, uniprot_get_proteome, uniprot_get_taxonomy, uniprot_get_sequence
• 2 resources: uniprot://entry/{accession}, uniprot://taxonomy/{taxonId}
• 1 prompt: uniprot_protein_dossier (guided protein-research workflow)
• UniProtService REST client over all four UniProt collections; UNIPROT_* config overrides
• Built on @cyanheads/mcp-ts-core — STDIO + Streamable HTTP, unified error handling, structured logging, optional OpenTelemetry

Changed:

• npm identity scoped to @cyanheads/uniprot-mcp-server across install surfaces; runtimeHint bun; tightened package/plugin/OCI descriptions, expanded keywords, added bugs/homepage/funding/repository metadata

102 tests pass; bun run devcheck clean.