@delta-infra/cli 1.1.17 → 1.1.19

package/skills/delta-protein/references/easifa_predict_active_site.md

@@ -0,0 +1,70 @@
+# Enzyme Active-site Prediction (`huada_protein_tools` / `easifa_predict_active_site`)
+
+This endpoint is a real endpoint of the `huada_protein_tools` tool. It can only be
+called via `delta-cli science invoke`; the business response described below lives
+under the CLI top-level `data` — read fields as documented in this file, do not
+recursively search. This file only documents the endpoint schema; it does not
+authorize direct HTTP. Endpoint names are authoritative as returned by
+`delta-cli science endpoints list huada_protein_tools`; use them verbatim — never
+guess, truncate, or alias.
+
+## Overview
+
+Predict enzyme active sites with EasIFA.
+- Required: `input_mode` (structure/sequence/uniprot)
+- structure mode requires `structure_file`; sequence mode requires `protein_sequence` or `sequence_file`
+- Optional: `rxn_smiles` (reaction SMILES), `chain`
+- Output: active-site predictions `resultjson`, `download_url`
+
+## Invocation
+
+```text
+delta-cli science invoke --tool huada_protein_tools --endpoint easifa_predict_active_site --data JSON
+```
+
+Required input groups (at least one per group; all groups must be satisfied):
+
+- `input_mode`
+
+Server-side default parameters:
+
+- `input_mode` = `structure`
+
+Request fields:
+
+| Field | Description | Default |
+|---|---|---|
+| `input_mode` | Input mode: structure/sequence/uniprot | structure |
+| `structure_file` | PDB/CIF | - |
+| `protein_sequence` |  | - |
+| `sequence_file` | FASTA/FA | - |
+| `rxn_smiles` | Reaction SMILES (substrate>>product) | - |
+| `chain` | Protein chain identifier, default A | A |
+
+## Response
+
+- CLI top-level envelope: a call succeeds only when the exit code is 0 AND `ok` is `true`; the service response lives under `data`.
+- Server-side contract: `status` (success/error/partial), `message`, `tool_type`, `resultjson`, `download_url`; some endpoints also return `total_count`. `resultjson` may be a single object or an array — read only the keys documented in this file, do not recursively search.
+- Binary/file artifacts (such as the PDB/XLSX/FASTA/ZIP files that `download_url` points to) are downloaded and verified with
+  `delta-cli science file download --tool huada_protein_tools --endpoint ... --output <path>`;
+  do not treat file contents as ordinary JSON results.
+
+## Output Projection
+
+> Verification status: derived from ToolSpec/agent templates; pending live calibration
+
+- Projection content: Enzyme active-site/binding-site residues, positions and confidence; positions overlapping zero-shot mutant results are marked as not recommended for engineering
+- Presentation form: Text/table
+- Downstream reuse: site information cross-annotated with zero_shot_mutant_candidates overlap
+## Output Template
+
+- Title: **Enzyme active-site/binding-site prediction result (EasIFA)**
+- Lead with the conclusion: active-site residues, positions, confidence
+- When zero-shot mutant results exist, mark overlapping positions (overlaps not recommended for engineering)
+
+## Boundaries & Failure Modes
+
+- File-type parameters (`structure_file`/`sequence_file`/`msa_file` etc.) take server-accessible URL/path strings, not local upload paths; a `download_url` or `pdb_url` returned by an upstream endpoint can be backfilled directly as downstream input.
+- Make only the minimal calls needed to complete the task; `--data` carries only the fields declared by this endpoint. Do not invent enum values or branches beyond the required combinations.
+- Do not pass off local tool (RDKit/pymatgen/LAMMPS/self-attached MCP) results as this endpoint's results; do not connect directly to Huada or gateway URLs.
+- On failure, report a short summary of the CLI's raw error; do not blindly retry after remote computation timeouts (the outcome may be unknown because the task may already have happened).

package/skills/delta-protein/references/ephod_predict_mutation.md

@@ -0,0 +1,66 @@
+# Mutant Optimal pH Prediction (`huada_protein_tools` / `ephod_predict_mutation`)
+
+This endpoint is a real endpoint of the `huada_protein_tools` tool. It can only be
+called via `delta-cli science invoke`; the business response described below lives
+under the CLI top-level `data` — read fields as documented in this file, do not
+recursively search. This file only documents the endpoint schema; it does not
+authorize direct HTTP. Endpoint names are authoritative as returned by
+`delta-cli science endpoints list huada_protein_tools`; use them verbatim — never
+guess, truncate, or alias.
+
+## Overview
+
+Predict the optimal pH of mutant proteins.
+- Input: `protein_sequence`/`sequence_file`/`structure_file` + `mutant_list`/`mutant_list_file`
+- Optional: `chain` (A), `target_pH` (-1.0 returns all)
+- Output: mutant optimal pH results `download_url`
+
+## Invocation
+
+```text
+delta-cli science invoke --tool huada_protein_tools --endpoint ephod_predict_mutation --data JSON
+```
+
+Required input groups (at least one per group; all groups must be satisfied):
+
+- `protein_sequence` / `sequence_file` / `structure_file`
+
+Request fields:
+
+| Field | Description | Default |
+|---|---|---|
+| `protein_sequence` |  | - |
+| `sequence_file` | FASTA/FA | - |
+| `structure_file` | PDB/CIF | - |
+| `chain` | Protein chain identifier, default A | A |
+| `mutant_list` | Mutation list, e.g. ["A1V", "K2M"] | [] |
+| `mutant_list_file` | CSV/XLSX mutation list file | - |
+| `target_pH` | Target pH, default -1.0 | -1.0 |
+
+## Response
+
+- CLI top-level envelope: a call succeeds only when the exit code is 0 AND `ok` is `true`; the service response lives under `data`.
+- Server-side contract: `status` (success/error/partial), `message`, `tool_type`, `resultjson`, `download_url`; some endpoints also return `total_count`. `resultjson` may be a single object or an array — read only the keys documented in this file, do not recursively search.
+- Binary/file artifacts (such as the PDB/XLSX/FASTA/ZIP files that `download_url` points to) are downloaded and verified with
+  `delta-cli science file download --tool huada_protein_tools --endpoint ... --output <path>`;
+  do not treat file contents as ordinary JSON results.
+
+## Output Projection
+
+> Verification status: derived from ToolSpec/agent templates; pending live calibration
+
+- Projection content: Optimal pH per mutant and its change (download_url)
+- Presentation form: Text/table
+- Downstream reuse: -
+## Output Template
+
+- Title: **Mutant optimal pH prediction result**
+- Lead with the conclusion: optimal pH per mutant and differences vs wild type
+- Add details to the "Result files & downloads" section
+
+## Boundaries & Failure Modes
+
+- File-type parameters (`structure_file`/`sequence_file`/`msa_file` etc.) take server-accessible URL/path strings, not local upload paths; a `download_url` or `pdb_url` returned by an upstream endpoint can be backfilled directly as downstream input.
+- Make only the minimal calls needed to complete the task; `--data` carries only the fields declared by this endpoint. Do not invent enum values or branches beyond the required combinations.
+- Do not pass off local tool (RDKit/pymatgen/LAMMPS/self-attached MCP) results as this endpoint's results; do not connect directly to Huada or gateway URLs.
+- On failure, report a short summary of the CLI's raw error; do not blindly retry after remote computation timeouts (the outcome may be unknown because the task may already have happened).

package/skills/delta-protein/references/ephod_predict_protein.md

@@ -0,0 +1,61 @@
+# Protein Optimal pH Prediction (`huada_protein_tools` / `ephod_predict_protein`)
+
+This endpoint is a real endpoint of the `huada_protein_tools` tool. It can only be
+called via `delta-cli science invoke`; the business response described below lives
+under the CLI top-level `data` — read fields as documented in this file, do not
+recursively search. This file only documents the endpoint schema; it does not
+authorize direct HTTP. Endpoint names are authoritative as returned by
+`delta-cli science endpoints list huada_protein_tools`; use them verbatim — never
+guess, truncate, or alias.
+
+## Overview
+
+Predict the optimal pH of proteins.
+- Input: `protein_list` (list of one-letter sequences) / `protein_list_file` (sequence file URL) — choose one
+- Important: `protein_list_file` must be a `.fasta/.fa/.csv/.xlsx` sequence-file URL; never pass `.pdb/.cif` structure files
+- Output: optimal pH prediction results `download_url`
+
+## Invocation
+
+```text
+delta-cli science invoke --tool huada_protein_tools --endpoint ephod_predict_protein --data JSON
+```
+
+Required input groups (at least one per group; all groups must be satisfied):
+
+- `protein_list` / `protein_list_file`
+
+Request fields:
+
+| Field | Description | Default |
+|---|---|---|
+| `protein_list` | List of single or multiple one-letter amino-acid sequences | [] |
+| `protein_list_file` | FASTA/FA/CSV/XLSX sequence file | - |
+
+## Response
+
+- CLI top-level envelope: a call succeeds only when the exit code is 0 AND `ok` is `true`; the service response lives under `data`.
+- Server-side contract: `status` (success/error/partial), `message`, `tool_type`, `resultjson`, `download_url`; some endpoints also return `total_count`. `resultjson` may be a single object or an array — read only the keys documented in this file, do not recursively search.
+- Binary/file artifacts (such as the PDB/XLSX/FASTA/ZIP files that `download_url` points to) are downloaded and verified with
+  `delta-cli science file download --tool huada_protein_tools --endpoint ... --output <path>`;
+  do not treat file contents as ordinary JSON results.
+
+## Output Projection
+
+> Verification status: derived from ToolSpec/agent templates; pending live calibration
+
+- Projection content: Optimal pH for each protein (download_url)
+- Presentation form: Text/table
+- Downstream reuse: -
+## Output Template
+
+- Title: **Protein optimal pH prediction result**
+- Lead with the conclusion: optimal pH range per protein
+- Add details to the "Result files & downloads" section
+
+## Boundaries & Failure Modes
+
+- File-type parameters (`structure_file`/`sequence_file`/`msa_file` etc.) take server-accessible URL/path strings, not local upload paths; a `download_url` or `pdb_url` returned by an upstream endpoint can be backfilled directly as downstream input.
+- Make only the minimal calls needed to complete the task; `--data` carries only the fields declared by this endpoint. Do not invent enum values or branches beyond the required combinations.
+- Do not pass off local tool (RDKit/pymatgen/LAMMPS/self-attached MCP) results as this endpoint's results; do not connect directly to Huada or gateway URLs.
+- On failure, report a short summary of the CLI's raw error; do not blindly retry after remote computation timeouts (the outcome may be unknown because the task may already have happened).

package/skills/delta-protein/references/esmfold2_predict.md

@@ -0,0 +1,67 @@
+# Protein Structure Prediction (ESMFold2) (`huada_protein_tools` / `esmfold2_predict`)
+
+This endpoint is a real endpoint of the `huada_protein_tools` tool. It can only be
+called via `delta-cli science invoke`; the business response described below lives
+under the CLI top-level `data` — read fields as documented in this file, do not
+recursively search. This file only documents the endpoint schema; it does not
+authorize direct HTTP. Endpoint names are authoritative as returned by
+`delta-cli science endpoints list huada_protein_tools`; use them verbatim — never
+guess, truncate, or alias.
+
+## Overview
+
+Predict the three-dimensional structure of protein/DNA/RNA/ligand complexes with
+ESMFold2; supports combined prediction.
+- Protein input: `protein_sequence` (sequence) / `sequence_file` (FASTA) — choose one
+- Optional: `msa_file` (A3M), `dna_sequence` (DNA sequence), `rna_sequence` (RNA sequence), `ligand` (ligand CCD code or SMILES)
+- Distinguish from Protenix: use ESMFold2 when DNA/RNA/ligand are involved; use Protenix for pure protein structure prediction
+- Output: predicted structure CIF file `download_url`, confidence metrics `resultjson`
+
+## Invocation
+
+```text
+delta-cli science invoke --tool huada_protein_tools --endpoint esmfold2_predict --data JSON
+```
+
+Required input groups (at least one per group; all groups must be satisfied):
+
+- `protein_sequence` / `sequence_file`
+
+Request fields:
+
+| Field | Description | Default |
+|---|---|---|
+| `protein_sequence` |  | - |
+| `sequence_file` | FASTA/FA | - |
+| `msa_file` | A3M MSA | - |
+| `dna_sequence` | DNA sequence, single str or list[str] | - |
+| `rna_sequence` | RNA sequence, single str or list[str] | - |
+| `ligand` | Small-molecule ligand: CCD code (e.g. ATP, MG) or SMILES string | - |
+
+## Response
+
+- CLI top-level envelope: a call succeeds only when the exit code is 0 AND `ok` is `true`; the service response lives under `data`.
+- Server-side contract: `status` (success/error/partial), `message`, `tool_type`, `resultjson`, `download_url`; some endpoints also return `total_count`. `resultjson` may be a single object or an array — read only the keys documented in this file, do not recursively search.
+- Binary/file artifacts (such as the PDB/XLSX/FASTA/ZIP files that `download_url` points to) are downloaded and verified with
+  `delta-cli science file download --tool huada_protein_tools --endpoint ... --output <path>`;
+  do not treat file contents as ordinary JSON results.
+
+## Output Projection
+
+> Verification status: verified by a real `delta-cli` invocation
+
+- Projection content: Complex structure download_url (.cif, composition when protein+DNA/RNA/ligand) with confidence: plddt_mean, ptm, iptm, sequence_length
+- Presentation form: Structure file link + confidence text
+- Downstream reuse: structure can serve as downstream structure_file
+## Output Template
+
+- Title: **ESMFold2 structure prediction result**
+- Lead with the conclusion: prediction done (complex composition, e.g. protein(47aa)+dna(12nt)) and confidence (plddt_mean/ptm/iptm), sequence_length
+- Add the structure CIF link to the "Result files & downloads" section
+
+## Boundaries & Failure Modes
+
+- File-type parameters (`structure_file`/`sequence_file`/`msa_file` etc.) take server-accessible URL/path strings, not local upload paths; a `download_url` or `pdb_url` returned by an upstream endpoint can be backfilled directly as downstream input.
+- Make only the minimal calls needed to complete the task; `--data` carries only the fields declared by this endpoint. Do not invent enum values or branches beyond the required combinations.
+- Do not pass off local tool (RDKit/pymatgen/LAMMPS/self-attached MCP) results as this endpoint's results; do not connect directly to Huada or gateway URLs.
+- On failure, report a short summary of the CLI's raw error; do not blindly retry after remote computation timeouts (the outcome may be unknown because the task may already have happened).

package/skills/delta-protein/references/evcouplings_mutation.md

@@ -0,0 +1,64 @@
+# EVcouplings Coevolution & Mutation Analysis (`huada_protein_tools` / `evcouplings_mutation`)
+
+This endpoint is a real endpoint of the `huada_protein_tools` tool. It can only be
+called via `delta-cli science invoke`; the business response described below lives
+under the CLI top-level `data` — read fields as documented in this file, do not
+recursively search. This file only documents the endpoint schema; it does not
+authorize direct HTTP. Endpoint names are authoritative as returned by
+`delta-cli science endpoints list huada_protein_tools`; use them verbatim — never
+guess, truncate, or alias.
+
+## Overview
+
+Infer a coevolutionary coupling model from an equal-length aligned MSA with plmc and
+output single-site mutation statistical energy scores plus coupling parameters.
+- Input: `msa_file` (equal-length aligned MSA file URL, FASTA/Clustal/Stockholm)
+- Optional: `query_id` (focus sequence ID in the MSA, default first), `model_file` (existing plmc .model file), `mutant_list_file` (Excel/CSV with a mutant column)
+- Output: single-site mutation score Excel `download_url`, `zip_download_url` (contains .model, coupling parameters, etc.)
+
+## Invocation
+
+```text
+delta-cli science invoke --tool huada_protein_tools --endpoint evcouplings_mutation --data JSON
+```
+
+Required input groups (at least one per group; all groups must be satisfied):
+
+- `msa_file` / `model_file`
+
+Request fields:
+
+| Field | Description | Default |
+|---|---|---|
+| `msa_file` | Equal-length aligned MSA file URL (FASTA/Clustal/Stockholm) | - |
+| `query_id` | Focus sequence ID in the MSA, default first | - |
+| `model_file` | Existing plmc .model file URL or path | - |
+| `mutant_list_file` | Excel/CSV URL or path with a mutant column | - |
+
+## Response
+
+- CLI top-level envelope: a call succeeds only when the exit code is 0 AND `ok` is `true`; the service response lives under `data`.
+- Server-side contract: `status` (success/error/partial), `message`, `tool_type`, `resultjson`, `download_url`; some endpoints also return `total_count`. `resultjson` may be a single object or an array — read only the keys documented in this file, do not recursively search.
+- Binary/file artifacts (such as the PDB/XLSX/FASTA/ZIP files that `download_url` points to) are downloaded and verified with
+  `delta-cli science file download --tool huada_protein_tools --endpoint ... --output <path>`;
+  do not treat file contents as ordinary JSON results.
+
+## Output Projection
+
+> Verification status: derived from ToolSpec/agent templates; pending live calibration
+
+- Projection content: Single-site mutation statistical energy scores (Excel download_url) + zip_download_url (.model and coupling parameters)
+- Presentation form: Energy score summary + Excel/ZIP links
+- Downstream reuse: -
+## Output Template
+
+- Title: **EVcouplings coevolutionary mutation analysis result**
+- Lead with the conclusion: TOP positions by single-site statistical energy score
+- Add the Excel and ZIP (with .model/coupling parameters) to the "Result files & downloads" section
+
+## Boundaries & Failure Modes
+
+- File-type parameters (`structure_file`/`sequence_file`/`msa_file` etc.) take server-accessible URL/path strings, not local upload paths; a `download_url` or `pdb_url` returned by an upstream endpoint can be backfilled directly as downstream input.
+- Make only the minimal calls needed to complete the task; `--data` carries only the fields declared by this endpoint. Do not invent enum values or branches beyond the required combinations.
+- Do not pass off local tool (RDKit/pymatgen/LAMMPS/self-attached MCP) results as this endpoint's results; do not connect directly to Huada or gateway URLs.
+- On failure, report a short summary of the CLI's raw error; do not blindly retry after remote computation timeouts (the outcome may be unknown because the task may already have happened).

package/skills/delta-protein/references/fasta_to_sequence.md

@@ -0,0 +1,59 @@
+# FASTA Sequence Extraction (`huada_protein_tools` / `fasta_to_sequence`)
+
+This endpoint is a real endpoint of the `huada_protein_tools` tool. It can only be
+called via `delta-cli science invoke`; the business response described below lives
+under the CLI top-level `data` — read fields as documented in this file, do not
+recursively search. This file only documents the endpoint schema; it does not
+authorize direct HTTP. Endpoint names are authoritative as returned by
+`delta-cli science endpoints list huada_protein_tools`; use them verbatim — never
+guess, truncate, or alias.
+
+## Overview
+
+Extract the protein sequence from a FASTA file.
+- Required: `fasta_file`
+- Output: extracted protein sequence `resultjson`
+
+## Invocation
+
+```text
+delta-cli science invoke --tool huada_protein_tools --endpoint fasta_to_sequence --data JSON
+```
+
+Required input groups (at least one per group; all groups must be satisfied):
+
+- `fasta_file`
+
+Request fields:
+
+| Field | Description | Default |
+|---|---|---|
+| `fasta_file` | FASTA/FA | - |
+
+## Response
+
+- CLI top-level envelope: a call succeeds only when the exit code is 0 AND `ok` is `true`; the service response lives under `data`.
+- Server-side contract: `status` (success/error/partial), `message`, `tool_type`, `resultjson`, `download_url`; some endpoints also return `total_count`. `resultjson` may be a single object or an array — read only the keys documented in this file, do not recursively search.
+- Binary/file artifacts (such as the PDB/XLSX/FASTA/ZIP files that `download_url` points to) are downloaded and verified with
+  `delta-cli science file download --tool huada_protein_tools --endpoint ... --output <path>`;
+  do not treat file contents as ordinary JSON results.
+
+## Output Projection
+
+> Verification status: derived from ToolSpec/agent templates; pending live calibration
+
+- Projection content: Protein sequence extracted from FASTA
+- Presentation form: Sequence shown as inline code
+- Downstream reuse: sequence can be used as downstream protein_sequence input
+## Output Template
+
+- Title: **FASTA sequence parsing result**
+- Lead with the conclusion: parsed sequence id/length
+- Show the sequence as inline code
+
+## Boundaries & Failure Modes
+
+- File-type parameters (`structure_file`/`sequence_file`/`msa_file` etc.) take server-accessible URL/path strings, not local upload paths; a `download_url` or `pdb_url` returned by an upstream endpoint can be backfilled directly as downstream input.
+- Make only the minimal calls needed to complete the task; `--data` carries only the fields declared by this endpoint. Do not invent enum values or branches beyond the required combinations.
+- Do not pass off local tool (RDKit/pymatgen/LAMMPS/self-attached MCP) results as this endpoint's results; do not connect directly to Huada or gateway URLs.
+- On failure, report a short summary of the CLI's raw error; do not blindly retry after remote computation timeouts (the outcome may be unknown because the task may already have happened).

package/skills/delta-protein/references/foldseek.md

@@ -0,0 +1,61 @@
+# Foldseek Sequence Extraction (`huada_protein_tools` / `foldseek`)
+
+This endpoint is a real endpoint of the `huada_protein_tools` tool. It can only be
+called via `delta-cli science invoke`; the business response described below lives
+under the CLI top-level `data` — read fields as documented in this file, do not
+recursively search. This file only documents the endpoint schema; it does not
+authorize direct HTTP. Endpoint names are authoritative as returned by
+`delta-cli science endpoints list huada_protein_tools`; use them verbatim — never
+guess, truncate, or alias.
+
+## Overview
+
+Extract the Foldseek/SaProt combined sequence from a protein structure file.
+- Required: `structure_file` (.pdb/.cif)
+- Optional: `chain` (A)
+- Output: extraction summary, `download_url` (FASTA)
+
+## Invocation
+
+```text
+delta-cli science invoke --tool huada_protein_tools --endpoint foldseek --data JSON
+```
+
+Required input groups (at least one per group; all groups must be satisfied):
+
+- `structure_file`
+
+Request fields:
+
+| Field | Description | Default |
+|---|---|---|
+| `structure_file` | PDB/CIF | - |
+| `chain` | Protein chain identifier, default A | A |
+
+## Response
+
+- CLI top-level envelope: a call succeeds only when the exit code is 0 AND `ok` is `true`; the service response lives under `data`.
+- Server-side contract: `status` (success/error/partial), `message`, `tool_type`, `resultjson`, `download_url`; some endpoints also return `total_count`. `resultjson` may be a single object or an array — read only the keys documented in this file, do not recursively search.
+- Binary/file artifacts (such as the PDB/XLSX/FASTA/ZIP files that `download_url` points to) are downloaded and verified with
+  `delta-cli science file download --tool huada_protein_tools --endpoint ... --output <path>`;
+  do not treat file contents as ordinary JSON results.
+
+## Output Projection
+
+> Verification status: derived from ToolSpec/agent templates; pending live calibration
+
+- Projection content: Extracted Foldseek/SaProt combined sequence, download_url (FASTA)
+- Presentation form: Sequence inline code + FASTA link
+- Downstream reuse: sequence can serve as downstream protein_sequence input
+## Output Template
+
+- Title: **Foldseek structure search result**
+- Lead with the conclusion: sequence extraction summary (length)
+- Add sequence inline code + FASTA link to the "Result files & downloads" section
+
+## Boundaries & Failure Modes
+
+- File-type parameters (`structure_file`/`sequence_file`/`msa_file` etc.) take server-accessible URL/path strings, not local upload paths; a `download_url` or `pdb_url` returned by an upstream endpoint can be backfilled directly as downstream input.
+- Make only the minimal calls needed to complete the task; `--data` carries only the fields declared by this endpoint. Do not invent enum values or branches beyond the required combinations.
+- Do not pass off local tool (RDKit/pymatgen/LAMMPS/self-attached MCP) results as this endpoint's results; do not connect directly to Huada or gateway URLs.
+- On failure, report a short summary of the CLI's raw error; do not blindly retry after remote computation timeouts (the outcome may be unknown because the task may already have happened).

package/skills/delta-protein/references/generate_msa.md

@@ -0,0 +1,61 @@
+# Multiple Sequence Alignment Computation (`huada_protein_tools` / `generate_msa`)
+
+This endpoint is a real endpoint of the `huada_protein_tools` tool. It can only be
+called via `delta-cli science invoke`; the business response described below lives
+under the CLI top-level `data` — read fields as documented in this file, do not
+recursively search. This file only documents the endpoint schema; it does not
+authorize direct HTTP. Endpoint names are authoritative as returned by
+`delta-cli science endpoints list huada_protein_tools`; use them verbatim — never
+guess, truncate, or alias.
+
+## Overview
+
+Compute a homolog protein MSA (A3M format) with MMseqs2 against the uniref100
+database.
+- Input: `protein_sequence` / `fasta_file` (choose one)
+- Output: A3M MSA file `download_url`
+
+## Invocation
+
+```text
+delta-cli science invoke --tool huada_protein_tools --endpoint generate_msa --data JSON
+```
+
+Required input groups (at least one per group; all groups must be satisfied):
+
+- `protein_sequence` / `fasta_file`
+
+Request fields:
+
+| Field | Description | Default |
+|---|---|---|
+| `protein_sequence` |  | - |
+| `fasta_file` | FASTA/FA | - |
+
+## Response
+
+- CLI top-level envelope: a call succeeds only when the exit code is 0 AND `ok` is `true`; the service response lives under `data`.
+- Server-side contract: `status` (success/error/partial), `message`, `tool_type`, `resultjson`, `download_url`; some endpoints also return `total_count`. `resultjson` may be a single object or an array — read only the keys documented in this file, do not recursively search.
+- Binary/file artifacts (such as the PDB/XLSX/FASTA/ZIP files that `download_url` points to) are downloaded and verified with
+  `delta-cli science file download --tool huada_protein_tools --endpoint ... --output <path>`;
+  do not treat file contents as ordinary JSON results.
+
+## Output Projection
+
+> Verification status: derived from ToolSpec/agent templates; pending live calibration
+
+- Projection content: Homolog MSA (A3M), download_url
+- Presentation form: Download link; note MSA can be used for homology/conservation inference
+- Downstream reuse: download_url -> msa_file for consensus_mutation / evcouplings_mutation / zero_shot_mutant_candidates
+## Output Template
+
+- Title: **Multiple sequence alignment (MSA) result**
+- Lead with the conclusion: MSA computed, usable for homology and conservation inference
+- Add the MSA file link to the "Result files & downloads" section and note it can serve as downstream msa_file
+
+## Boundaries & Failure Modes
+
+- File-type parameters (`structure_file`/`sequence_file`/`msa_file` etc.) take server-accessible URL/path strings, not local upload paths; a `download_url` or `pdb_url` returned by an upstream endpoint can be backfilled directly as downstream input.
+- Make only the minimal calls needed to complete the task; `--data` carries only the fields declared by this endpoint. Do not invent enum values or branches beyond the required combinations.
+- Do not pass off local tool (RDKit/pymatgen/LAMMPS/self-attached MCP) results as this endpoint's results; do not connect directly to Huada or gateway URLs.
+- On failure, report a short summary of the CLI's raw error; do not blindly retry after remote computation timeouts (the outcome may be unknown because the task may already have happened).

package/skills/delta-protein/references/interpro_analyze.md

@@ -0,0 +1,63 @@
+# InterProScan Domain & Function Annotation (`huada_protein_tools` / `interpro_analyze`)
+
+This endpoint is a real endpoint of the `huada_protein_tools` tool. It can only be
+called via `delta-cli science invoke`; the business response described below lives
+under the CLI top-level `data` — read fields as documented in this file, do not
+recursively search. This file only documents the endpoint schema; it does not
+authorize direct HTTP. Endpoint names are authoritative as returned by
+`delta-cli science endpoints list huada_protein_tools`; use them verbatim — never
+guess, truncate, or alias.
+
+## Overview
+
+Annotate protein domains and functions with InterProScan.
+- Input: `protein_sequence` / `sequence_file` (FASTA/FA) / `structure_file` (PDB/CIF) — choose one; when `structure_file` is provided the backend extracts the sequence automatically
+- Optional: `chain` (protein chain identifier, default A)
+- Output: domain/family/GO summary `resultjson`, Excel (.xlsx) download link `download_url`
+
+## Invocation
+
+```text
+delta-cli science invoke --tool huada_protein_tools --endpoint interpro_analyze --data JSON
+```
+
+Required input groups (at least one per group; all groups must be satisfied):
+
+- `protein_sequence` / `sequence_file` / `structure_file`
+
+Request fields:
+
+| Field | Description | Default |
+|---|---|---|
+| `protein_sequence` |  | - |
+| `sequence_file` | FASTA/FA | - |
+| `structure_file` | PDB/CIF | - |
+| `chain` | Protein chain identifier, default A | A |
+
+## Response
+
+- CLI top-level envelope: a call succeeds only when the exit code is 0 AND `ok` is `true`; the service response lives under `data`.
+- Server-side contract: `status` (success/error/partial), `message`, `tool_type`, `resultjson`, `download_url`; some endpoints also return `total_count`. `resultjson` may be a single object or an array — read only the keys documented in this file, do not recursively search.
+- Binary/file artifacts (such as the PDB/XLSX/FASTA/ZIP files that `download_url` points to) are downloaded and verified with
+  `delta-cli science file download --tool huada_protein_tools --endpoint ... --output <path>`;
+  do not treat file contents as ordinary JSON results.
+
+## Output Projection
+
+> Verification status: derived from ToolSpec/agent templates; pending live calibration
+
+- Projection content: Domain, family, GO annotation summary + Excel (download_url)
+- Presentation form: Annotation summary text + Excel download link
+- Downstream reuse: -
+## Output Template
+
+- Title: **InterProScan domain and function annotation result**
+- Lead with the conclusion: key domain/family/GO annotation entries
+- Add the Excel detail file to the "Result files & downloads" section
+
+## Boundaries & Failure Modes
+
+- File-type parameters (`structure_file`/`sequence_file`/`msa_file` etc.) take server-accessible URL/path strings, not local upload paths; a `download_url` or `pdb_url` returned by an upstream endpoint can be backfilled directly as downstream input.
+- Make only the minimal calls needed to complete the task; `--data` carries only the fields declared by this endpoint. Do not invent enum values or branches beyond the required combinations.
+- Do not pass off local tool (RDKit/pymatgen/LAMMPS/self-attached MCP) results as this endpoint's results; do not connect directly to Huada or gateway URLs.
+- On failure, report a short summary of the CLI's raw error; do not blindly retry after remote computation timeouts (the outcome may be unknown because the task may already have happened).

package/skills/delta-protein/references/mpnn_design_sequence.md

@@ -0,0 +1,71 @@
+# MPNN Sequence Design (`huada_protein_tools` / `mpnn_design_sequence`)
+
+This endpoint is a real endpoint of the `huada_protein_tools` tool. It can only be
+called via `delta-cli science invoke`; the business response described below lives
+under the CLI top-level `data` — read fields as documented in this file, do not
+recursively search. This file only documents the endpoint schema; it does not
+authorize direct HTTP. Endpoint names are authoritative as returned by
+`delta-cli science endpoints list huada_protein_tools`; use them verbatim — never
+guess, truncate, or alias.
+
+## Overview
+
+Protein sequence design with the ProteinMPNN family of models.
+- Required: `structure_file` (.pdb/.cif)
+- Optional: `model_type` (protein_mpnn/ligand_mpnn/soluble_mpnn), `number` (1), `fixed_residues`, `redesigned_residues`, `chain`
+- Output: designed sequences with scores `resultjson`, `download_url`
+- Use case: when only a PDB/CIF structure file is provided
+
+## Invocation
+
+```text
+delta-cli science invoke --tool huada_protein_tools --endpoint mpnn_design_sequence --data JSON
+```
+
+Required input groups (at least one per group; all groups must be satisfied):
+
+- `structure_file`
+
+Server-side default parameters:
+
+- `model_type` = `protein_mpnn`
+- `number` = `1`
+
+Request fields:
+
+| Field | Description | Default |
+|---|---|---|
+| `structure_file` | PDB/CIF | - |
+| `model_type` | Model type, default protein_mpnn | protein_mpnn |
+| `number` | Number of sequences to generate, default 1 | 1 |
+| `fixed_residues` | Fixed residue positions, space-separated | - |
+| `redesigned_residues` | Positions to redesign only, space-separated | - |
+| `chain` | PDB chain identifier | - |
+
+## Response
+
+- CLI top-level envelope: a call succeeds only when the exit code is 0 AND `ok` is `true`; the service response lives under `data`.
+- Server-side contract: `status` (success/error/partial), `message`, `tool_type`, `resultjson`, `download_url`; some endpoints also return `total_count`. `resultjson` may be a single object or an array — read only the keys documented in this file, do not recursively search.
+- Binary/file artifacts (such as the PDB/XLSX/FASTA/ZIP files that `download_url` points to) are downloaded and verified with
+  `delta-cli science file download --tool huada_protein_tools --endpoint ... --output <path>`;
+  do not treat file contents as ordinary JSON results.
+
+## Output Projection
+
+> Verification status: derived from ToolSpec/agent templates; pending live calibration
+
+- Projection content: Number of designed sequences, scores/confidence, download_url
+- Presentation form: Text + download link
+- Downstream reuse: -
+## Output Template
+
+- Title: **MPNN sequence design result**
+- Lead with the conclusion: number of designed sequences and scores/confidence
+- Add the design-result file link to the "Result files & downloads" section
+
+## Boundaries & Failure Modes
+
+- File-type parameters (`structure_file`/`sequence_file`/`msa_file` etc.) take server-accessible URL/path strings, not local upload paths; a `download_url` or `pdb_url` returned by an upstream endpoint can be backfilled directly as downstream input.
+- Make only the minimal calls needed to complete the task; `--data` carries only the fields declared by this endpoint. Do not invent enum values or branches beyond the required combinations.
+- Do not pass off local tool (RDKit/pymatgen/LAMMPS/self-attached MCP) results as this endpoint's results; do not connect directly to Huada or gateway URLs.
+- On failure, report a short summary of the CLI's raw error; do not blindly retry after remote computation timeouts (the outcome may be unknown because the task may already have happened).