MCP server for FactorForge CDS design and bioinformatics lookup tools. Provides in-silico DNA sequence optimization, protein structure queries, literature search, and pathway analysis.
22 tools with complete JSON schemas and descriptions, but significant gaps in parameter documentation and error handling. Most tools have 1 - 3 parameters with basic type info but lack actionable constraints (enums, ranges, format hints). Descriptions are present (avg ~100 chars) but often generic or research-focused rather than LLM-optimized. No output schemas documented. Error recovery guidance absent. Tool naming is mostly clear (verb_noun pattern) but some names are domain-specific jargon that may confuse agents unfamiliar with synthetic biology (e.g., 'factorforge_cds_optimize', 'prd_b_*'). Composition is reasonable, tools are single-purpose, but parameter relationships and dependencies are underdocumented.
[Production] Evaluate DNA against AgentOS Hard Gate.
Generate CDS candidates for up to 20 protein sequences in a single request.
Compare multiple FactorForge CDS design profiles side-by-side.
Generate an in-silico synonymous CDS candidate for N. benthamiana.
Generate a bounded Top-K computational discovery slate.
Initialize a structured 0→7 step research workflow to verify or update a FactorForge design constant (e.g. GC_OPT_MIN, CAI_THRESHOLD). Returns a ready-to-execute research plan with pre-filled PubMed queries and decision gates.
No output schemas documented. Tools return results but LLMs cannot plan downstream calls or extract fields without trial-and-error. Violates pattern:tool and pattern:response-shaper.
PRD-B tool names use opaque prefixes ('prd_b_*') that obscure intent. LLMs cannot infer what PRD-B is or when to use these tools. Naming violates pattern:tool and mxe:chat-data-model.
Parameter descriptions lack actionable constraints. 'top_k: 1 - 10; default 3' is stated in description but not enforced as JSON Schema minItems/maxItems. Numeric ranges, enums, and format hints must be in schema, not prose. Violates pattern:constrained-input.
Inferred effective spec: <=2025-11-25.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | C | 60 | <=2025-11-25 | v2 |
[Research Preview] FactorForge v3.5 sLLM Hybrid (Gen 3).
Commit a private design record to PRD-B.
Flush a design record to public projection (irreversible).
Generate an approval packet for a design record.
Get the approval status of a design record.
Retrieve a private design record from PRD-B.
Retrieve a public design record projection.
List PRD-B index entries.
Run FactorForge CDS design with PRD-B provenance tracking.
Search AlphaFold DB for predicted protein structures.
Search InterPro for protein domains and functional annotations.
Search KEGG for biological pathways.
Search NCBI protein or nucleotide database and return sequence summaries and accession numbers.
Search the RCSB Protein Data Bank for 3D protein structures.
Search PubMed for scientific literature.
Search UniProt for protein entries — accession, gene names, organism, function, and sequence length.
No error handling or recovery guidance. Tools do not document what errors are retryable, user-fixable, or fatal. Violates pattern:recovery-guide and pattern:error-classification.
Irreversible operation 'prd_b_flush_public_projection' lacks confirmation or dry-run support. Agents can accidentally publish records without safeguards. Violates pattern:confirmation-request.