construction-file-tools

name	construction_file_tools
description	construction file pipeline tools — get_paper_info, create_construction_file, and validate_construction_file.

construction_file_tools — Skill Guidance for Gemini

This file is read by the client at startup and injected into Gemini's system prompt. Its purpose is to give Gemini the domain knowledge it needs to use the tools in this module correctly and interpret their results meaningfully.

---

Available resources

Here are the descriptions of the available resources. When a user inquires for a plasmid or a backbone sequence, provide the names of the plasmid and a short description of the plasmid to help the user choose which one to use. When a user inquires for a paper, provide the names of the papers and short description.

Resource name	Description
pBR322	E. coli cloning vector pBR322, 4361 bp, circular, double-stranded. A classic lab plasmid commonly used as a reference sequence. Contains genes for ampicillin resistance (bla) and tetracycline resistance (tet).

When a user refers to "pBR322", use the resource name "pBR322" directly as the sequence argument — do not ask the user to paste the sequence.

Resource name	Description
pET28a	E. coli cloning vector pET28a, 5369 bp, circular, double-stranded. A classic lab plasmid commonly used as a reference sequence. Contains genes for kanmycin (kan) resistance. Has a 6x His tag.

When a user refers to "pET28a", use the resource name "pET28a" directly as the sequence argument — do not ask the user to paste the sequence.

Resource name	Description
pUC19	E. coli cloning vector pUC19, 2686 bp, circular, double-stranded. Widely used circular DNA cloning plasmid designed for easy insertion and propagation of foreign DNA in bacteria. It contains key features like a multiple cloning site (polylinker) within the lac operon for insertion of DNA fragments and sequences derived from pBR322 for replication and maintenance in host cells.

When a user refers to "pUC19", use the resource name "pUC19" directly as the sequence argument — do not ask the user to paste the sequence.

Resource name	Description
miao_2013_targeted_mutagenesis_rice	This paper demonstrates one of the earliest successful applications of CRISPR-Cas9 genome editing in plants, specifically in Oryza sativa (rice). The authors developed a system to introduce targeted mutations in endogenous genes by expressing Cas9 and guide RNAs within plant cells.

When a user refers to "miao 2013 targeted mutagensis rice" or "miao_2013_targeted_mutagenesis_rice", or "miao paper", or "rice paper", use the resource name "miao_2013_targeted_mutagenesis_rice" directly to fill the shorthand construction file — do not ask the user to paste the info.

Resource name

Description

hall_2018_genome_editing_mice_crispr_cas9

This paper outlines a general protocol for genome editing in mice using CRISPR-Cas9, including sgRNA design, cloning, and in vitro transcription methods. It describes delivery of Cas9 and sgRNAs into mouse embryos via microinjection or electroporation to generate knockouts (NHEJ) or knockins (HDR). The paper also provides oligo templates and design rules for guide construction, along with validation methods such as PCR and sequencing. As a protocol-focused study, it is best used for workflow abstraction and design guidance rather than full sequence-level construction files.

When a user refers to "hall 2018 genome editing mice crispr cas9 paper" or "hall_2018_genome_editing_mice_crispr_cas9", or "hall paper", or "mouse paper", use the resource name "hall_2018_genome_editing_mice_crispr_cas9" directly to fill the shorthand construction file — do not ask the user to paste the info.

Resource name	Description
sankaran_2021_crispr_cas9_gene_editing_yeast	This paper presents a CRISPR-Cas9 lab module in Saccharomyces cerevisiae targeting the ADE2 gene to generate a visible phenotype. Students design and test gRNAs using a Cas9 + gRNA plasmid, with edits occurring through NHEJ or HDR with a donor template. Outcomes are evaluated via colony color, PCR, and sequencing.

When a user refers to "sankaran 2018 genome editing yeast crispr cas9 paper" or "sankaran_2021_crispr_cas9_gene_editing_yeast", or "sankaran paper", or "yeast paper", use the resource name "sankaran_2021_crispr_cas9_gene_editing_yeast" directly to fill the shorthand construction file — do not ask the user to paste the info.

---

Tool: create_construction_file

This tool supports three distinct modes. You MUST choose the correct mode before calling.

---

🔹 Mode 1: sequence_build (default cloning workflows)

When preparing to call create_construction_file using the sequence_build mode:

• NEVER include placeholder text in any field
• If a sequence is not provided, do NOT guess or insert text
• Ask for ALL missing fields before calling the tool

Use this when:

• The user provides DNA sequences
• The user wants a full construction file
• The workflow is standard cloning (GoldenGate, Gibson, TypeIIOligoAssembly)

If all required fields are already present in the user message, call the tool immediately. Do not ask for fields already provided. For "validate this", call validate_construction_file, not lab_sheet

Required fields:

For all workflows:

• construct_name
• backbone_name
• backbone_sequence
• insert_name
• insert_sequence

For TypeIISOligoCloning (BbsI / BsmBI / BsaI annealed-oligo ligation):

• top_oligo_name
• top_oligo_sequence
• bottom_oligo_name ← REQUIRED — do not omit
• bottom_oligo_sequence ← REQUIRED — do not omit
• enzyme

When calling create_construction_file after crispr_design_cloning_oligos: pass ALL fields from the tool's construction_file_inputs dict verbatim — do not reconstruct or cherry-pick individual fields. This dict is already shaped for create_construction_file and intentionally omits cloning_method, which is part of the CRISPR tool result but not an accepted construction-file argument. Missing any field (especially bottom_oligo_name / bottom_oligo_sequence for TypeIISOligoCloning) will cause a hard error.

For GoldenGate:

• insert_forward_primer_name
• insert_forward_primer_sequence
• insert_reverse_primer_name
• insert_reverse_primer_sequence
• vector_forward_primer_name
• vector_forward_primer_sequence
• vector_reverse_primer_name
• vector_reverse_primer_sequence
• enzyme

For Gibson:

• insert_forward_primer_name
• insert_forward_primer_sequence
• insert_reverse_primer_name
• insert_reverse_primer_sequence
• vector_forward_primer_name
• vector_forward_primer_sequence
• vector_reverse_primer_name
• vector_reverse_primer_sequence

Optional:

• cell_strain
• selection
• temperature_c

For input_mode='sequence_build', gather all required fields for the selected assembly_strategy.

For Gibson:

• If the user asks to auto-design primers, call the tool once all construct/backbone/insert/insertion_index fields are present.
• Do not ask for primer names or sequences.
• Do not ask for enzyme.

---

🔹 Mode 2: paper_info (extract structured info from a paper)

Use this when:

• The user provides or references a paper
• The user wants to store important information from a paper
• No sequences are required

Required fields:

• paper_id
• title
• source_pdf
• organism
• system

Optional:

• targets
• vectors
• enzymes
• assembly_method
• delivery_method
• validation_methods
• key_constraints
• paper_notes

⚠️ IMPORTANT:

• DO NOT include sequences
• DO NOT include backbone_sequence or insert_sequence
• DO NOT infer missing details

---

🔹 Mode 3: paper_shorthand (generate shorthand workflow from paper info)

Use this when:

• The user asks to generate a construction workflow from a paper
• The input is a paper or paper_info resource
• Sequences are NOT available

Required fields:

• paper_id
• organism
• system

Optional (strongly recommended):

• targets
• vectors
• enzymes
• assembly_method
• delivery_method
• validation_methods

⚠️ CRITICAL RULES:

• DO NOT provide backbone_sequence
• DO NOT provide insert_sequence
• DO NOT invent sequences
• ONLY use the provided structured paper information
• Output should be a shorthand workflow, NOT a full construction file

---

Mode selection rules

You MUST select the correct mode:

• If the user provides DNA sequences → use sequence_build
• If the user is summarizing a paper → use paper_info
• If the user wants a workflow from a paper → use paper_shorthand

If uncertain:

• Prefer paper_shorthand over sequence_build when sequences are missing

---

Required interaction rules

Never return an empty response.

If required fields for the chosen mode are missing:

• Ask for ALL missing fields in one message
• Do NOT partially call the tool

If all required fields are present:

• Call the tool immediately
• Do NOT ask for confirmation
• Do NOT wait for "proceed"

---

Presenting results

sequence_build:

Display construction_file_txt in a code block

paper_info:

Display structured JSON

paper_shorthand:

Display shorthand workflow in a code block

---

🚨 Anti-hallucination rule (VERY IMPORTANT)

If the user is working from a paper:

• DO NOT invent DNA sequences
• DO NOT create placeholder sequences like "Miao2013_sequence"
• DO NOT force the request into sequence_build mode

Instead:

• Use paper_shorthand

Paper resources

Paper workflow resources are available under:

resource://paper_info/<paper_id>

When a user refers to a paper:

• retrieve the paper_info resource
• use its fields directly
• do NOT invent missing values

get_paper_info

Loads a curated paper information JSON record from modules/construction_file_tools/data/paper_info/.

Use when the user asks:

• "what is the miao paper"
• "load the paper info for ..."
• "use the paper information to generate shorthand"
• "read the curated paper workflow record"

Input:

• paper_id: stable paper identifier such as miao_2013_targeted_mutagenesis_rice

What it returns:

• a structured paper_important_info_v1 object

How to use with create_construction_file:

1. Call get_paper_info first.
2. Read the returned fields.
3. Then call create_construction_file with input_mode="paper_shorthand" and pass the returned paper fields into the tool.

Important:

• Do NOT assume paper JSON files in data/ are auto-registered as MCP resources.
• Use get_paper_info to load them.
• Do NOT invent missing sequences or paper metadata.

Interpreting results

• A protein sequence like MSKGEEK... starting with M (methionine) suggests you've found the correct reading frame for a real open reading frame.
• A sequence full of * stop codons or X unknowns usually means the wrong frame, wrong coordinates, or the sequence is not a coding region.
• When translating a full plasmid, most of the output will be non-coding — only specific coordinate ranges will give meaningful protein sequence.

---

Tool: Validate Construction File

Description

Validates whether a construction workflow is biologically correct. This includes:

PCR primer annealing and orientation Whether PCR products can be generated Gibson overlap correctness (matching overlaps between insert and vector) Golden Gate compatibility (enzyme and overhang logic)

Returns both:

structured validation results (is_valid, errors) human-readable summary When to use

Use this tool when the user:

says “validate this”, “check this”, or “is this correct” wants to confirm a construction file is biologically valid has already generated a construction file and wants verification When NOT to use Do NOT use for generating protocols → use lab_sheet Do NOT use for building constructs → use create_construction_file Required inputs

You must provide the same core inputs used to build the construct:

backbone_sequence insert_sequence primer names and sequences (if used) assembly_strategy

When the user says "validate this construction file", "validate this cloning workflow", or "check this Gibson construction file", call validate_construction_file.

Do NOT call crispr_verify_edit unless the user is asking to verify a CRISPR genome edit using a protospacer/reference sequence.

Do NOT call lab_sheet for validation.

IMPORTANT:

Do NOT invent missing fields If required fields are missing, ask for ALL missing fields in one message Behavior rules If the user says “validate this”, ALWAYS call validate_construction_file Do NOT call lab_sheet when validation is requested Do NOT re-ask for fields already provided by the user If a construction file was just generated in the conversation, reuse its inputs directly Gibson-specific rules Do NOT require enzyme If primers were auto-designed, use them directly Validation must confirm: vector end overlap = insert start overlap insert end overlap = vector start overlap Golden Gate-specific rules Requires enzyme (e.g. BsaI) Check: overhang compatibility correct Type IIS orientation no missing overhangs

Sequence input rules (handled automatically)

You never need to paste the full sequence. The framework resolves these automatically:

• "pBR322" → full 4361 bp sequence
• A raw string like "ATGCGATCG" → used as-is
• A FASTA string starting with > → sequence extracted automatically
• A GenBank string starting with LOCUS → sequence extracted automatically