Tooluniverse Target ResearchSAFE

The largest open-source medical AI skills library for OpenClaw🦞.

The instruction file, verbatim from the audited commit — this is the text the model reads, and the surface the audit's instruction layer examines. Quoted here so you can judge it without cloning anything.

---
name: tooluniverse-target-research
description: Gather comprehensive biological target intelligence from 9 parallel research paths covering protein info, structure, interactions, pathways, expression, variants, drug interactions, and literature. Features collision-aware searches, evidence grading (T1-T4), explicit Open Targets coverage, and mandatory completeness auditing. Use when users ask about drug targets, proteins, genes, or need target validation, druggability assessment, or comprehensive target profiling.
---

# Comprehensive Target Intelligence Gatherer

Gather complete target intelligence by exploring 9 parallel research paths. Supports targets identified by gene symbol, UniProt accession, Ensembl ID, or gene name.

**KEY PRINCIPLES**:
1. **Report-first approach** - Create report file FIRST, then populate progressively
2. **Tool parameter verification** - Verify params via `get_tool_info` before calling unfamiliar tools
3. **Evidence grading** - Grade all claims by evidence strength (T1-T4)
4. **Citation requirements** - Every fact must have inline source attribution
5. **Mandatory completeness** - All sections must exist with data minimums or explicit "No data" notes
6. **Disambiguation first** - Resolve all identifiers before research
7. **Negative results documented** - "No drugs found" is data; empty sections are failures
8. **Collision-aware literature search** - Detect and filter naming collisions
9. **English-first queries** - Always use English terms in tool calls, even if the user writes in another language. Translate gene names, disease names, and search terms to English. Only try original-language terms as a fallback if English returns no results. Respond in the user's language

---

## Phase 0: Tool Parameter Verification (CRITICAL)

**BEFORE calling ANY tool for the first time**, verify its parameters:

```python
# Always check tool params to prevent silent failures
tool_info = tu.tools.get_tool_info(tool_name="Reactome_map_uniprot_to_pathways")
# Reveals: takes `id` not `uniprot_id`
```

### Known Parameter Corrections (Updated)

| Tool | WRONG Parameter | CORRECT Parameter |
|------|-----------------|-------------------|
| `Reactome_map_uniprot_to_pathways` | `uniprot_id` | `id` |
| `ensembl_get_xrefs` | `gene_id` | `id` |
| `GTEx_get_median_gene_expression` | `gencode_id` only | `gencode_id` + `operation="median"` |
| `OpenTargets_*` | `ensemblID` | `ensemblId` (camelCase) |

### GTEx Versioned ID Fallback (CRITICAL)

GTEx often requires versioned Ensembl IDs. If `ENSG00000123456` returns empty:

```python
# Step 1: Get gene info with version
gene_info = tu.tools.ensembl_lookup_gene(id=ensembl_id, species="human")
version = gene_info.get('version', 1)

# Step 2: Try versioned ID
versioned_id = f"{ensembl_id}.{version}"  # e.g., "ENSG00000123456.12"
result = tu.tools.GTEx_get_median_gene_expression(
    gencode_id=versioned_id,
    operation="median"
)
```

---

## When to Use This Skill

Apply when users:
- Ask about a drug target, protein, or gene
- Need target validation or assessment
- Request druggability analysis
- Want comprehensive target profiling
- Ask "what do we know about [target]?"
- Need target-disease associations
- Request safety profile for a target

---

## Critical Workflow Requirements

### 1. Report-First Approach (MANDATORY)

**DO NOT** show the search process or tool outputs to the user. Instead:

1. **Create the report file FIRST** - Before any data collection:
   - File name: `[TARGET]_target_report.md`
   - Initialize with all 14 section headers
   - Add placeholder: `[Researching...]` in each section

2. **Progressively update the report** - As you gather data:
   - Update each section immediately after retrieving data
   - Replace `[Researching...]` with actual content
   - Include "No data returned" when tools return empty results

3. **Methodology in appendix only** - If user requests methodology details, create separate `[TARGET]_methods_appendix.md`

### 2. Evidence Grading System (MANDATORY)

**CRITICAL**: Grade every claim by evidence strength.

#### Evidence Tiers

| Tier | Symbol | Criteria | Examples |
|------|--------|----------|----------|
| **T1** | ★★★ | Direct mechanistic evidence, human genetic proof | CRISPR KO, patient mutations, crystal structure with mechanism |
| **T2** | ★★☆ | Functional studies, model organism validation | siRNA phenotype, mouse KO, biochemical assay |
| **T3** | ★☆☆ | Association, screen hits, computational | GWAS hit, DepMap essentiality, expression correlation |
| **T4** | ☆☆☆ | Mention, review, text-mined, predicted | Review article, database annotation, computational prediction |

#### Required Evidence Grading Locations

Evidence grades MUST appear in:
1. **Executive Summary** - Key disease claims graded
2. **Section 8.2 Disease Associations** - Every disease link graded with source type
3. **Section 11 Literature** - Key papers table with evidence tier
4. **Section 13 Recommendations** - Scorecard items reference evidence quality

#### Per-Section Evidence Summary

```markdown
---
**Evidence Quality for this Section**: Strong
- Mechanistic (T1): 12 papers
- Functional (T2): 8 papers
- Association (T3): 15 papers
- Mention (T4): 23 papers
**Data Gaps**: No CRISPR data; mouse KO phenotypes limited
---
```

### 3. Citation Requirements (MANDATORY)

Every piece of information MUST include its source:

```markdown
EGFR mutations cause lung adenocarcinoma [★★★: PMID:15118125, activating mutations 
in patients]. *Source: ClinVar, CIViC*
```

---

## Core Strategy: 9 Research Paths

Execute 9 research paths (Path 0 is always first):

```
Target Query (e.g., "EGFR" or "P00533")
│
├─ IDENTIFIER RESOLUTION (always first)
│   └─ Check if GPCR → GPCRdb_get_protein
│
├─ PATH 0: Open Targets Foundation (ALWAYS FIRST - fills gaps in all other paths)
│
├─ PATH 1: Core Identity (names, IDs, sequence, organism)
│   └─ InterProScan_scan_sequence for novel domain prediction (NEW)
├─ PATH 2: Structure & Dom

SAFEgrade B · trust 89/100 Nothing in the source contradicts what it says it does. Grade A is reserved for packages that have also passed the behavioural sandbox.

What the source does

Filesystem

none-observed

Network

none-observed

Shell

none-observed

Dependencies

pinned

Secrets in source

none-found

Findings (0)

No findings outside the package's declared scope.

Gates applied: no_behavioural_pass.