Skip to main content
skill CX Codex

tooluniverse-adverse-outcome-pathway — Claude Code skill setup

tooluniverse-adverse-outcome-pathway maps industrial and environmental chemicals to adverse outcome pathways. Retrieves GHS classifications, IARC cancer statuses, LD50 toxicity values, and CTD gene interactions.

Apache-2.0 236 installs ★ 1,798 updated 2026-10-24

Install into Claude Code · Skills CLI (Vercel) All methods and platforms ↓
npx skills add mims-harvard/ToolUniverse --skill tooluniverse-adverse-outcome-pathway -a claude-code
Open on GitHub

What tooluniverse-adverse-outcome-pathway does

This skill conducts mechanistic toxicology reviews for environmental toxicants and industrial chemicals. It tracks biological perturbations from molecular initiating events down to apical organ toxicity by querying AOPWiki, PubChemTox, and CTD.

The workflow requires obtaining a PubChem CID via PubChem_get_CID_by_compound_name before running hazard queries. CTD tools accept chemical names directly, while PubChemTox calls fetch GHS pictograms, IARC group classifications, and acute LD50 values across test species.

Do not point this skill at approved pharmaceutical drugs: it lacks FDA FAERS integration and is strictly built for environmental chemical risk assessment. For approved medications, use tooluniverse-toxicology.

What it helps with

  • map environmental toxicants to adverse outcome pathways in AOPWiki
  • retrieve compound GHS hazard categories and warning pictograms
  • extract species-specific LD50 and LC50 toxicity values
  • check compound carcinogenicity under IARC, NTP, and EPA classifications
  • inspect chemical-gene interactions and disease links in CTD

Example prompts for the agent

  • Check the AOPs, IARC status, and GHS classification for benzo[a]pyrene
  • Get the oral rat LD50 and acute symptoms for rotenone from PubChem
  • Find chemical-gene interactions and curated diseases for bisphenol A in CTD

How to install tooluniverse-adverse-outcome-pathway

1Skills CLI (Vercel)

npx skills add mims-harvard/ToolUniverse --skill tooluniverse-adverse-outcome-pathway -a claude-code

2Manually: put the files in place

~/.claude/skills/tooluniverse-adverse-outcome-pathway/SKILL.md

Copy the whole skill folder; for one project use .claude/skills/tooluniverse-adverse-outcome-pathway/

Step by step, with commands for each platform: How to install skills in Claude Code: five ways

Questions about tooluniverse-adverse-outcome-pathway

How to install tooluniverse-adverse-outcome-pathway in Claude Code?

Run npx skills add mims-harvard/ToolUniverse --skill tooluniverse-adverse-outcome-pathway -a claude-code. The skill file will be stored at ~/.claude/skills/tooluniverse-adverse-outcome-pathway/SKILL.md.

What does tooluniverse-adverse-outcome-pathway do?

It links chemical compounds to AOPWiki pathways, PubChem hazard classifications, and CTD toxicogenomics data to evaluate non-drug chemical toxicity.

Can I use this skill to check adverse effects of pharmaceutical drugs?

No. It does not parse FAERS data or FDA drug labels. For approved prescription drugs, the author specifies using tooluniverse-toxicology.

Source text

SKILL.md · Apache-2.0
---
name: tooluniverse-adverse-outcome-pathway
description: Map environmental and industrial chemicals to adverse outcome pathways (AOPs) — molecular initiating event to organ-level toxicity. Uses AOPWiki, GHS classification, IARC carcinogen status, and LD50 data. Use for environmental/industrial chemical risk assessment, regulatory-grade hazard characterization, and AOP stressor mapping. Distinct from drug-safety analysis (use tooluniverse-pharmacovigilance for drugs).
disable-model-invocation: true
---

# Adverse Outcome Pathway & Regulatory Risk Assessment

Distinct from drug safety (see tooluniverse-toxicology): this skill targets **environmental and
industrial chemicals** where the focus is AOP stressor mapping, GHS classification, LD50 hazard
quantification, and IARC carcinogen status — not FAERS signals or FDA drug labels.

## When to Use

Apply when researcher asks about:
- "What AOPs are associated with [pesticide/solvent/industrial chemical]?"
- "What is the GHS hazard classification for [compound]?"
- "What is the LD50 for [compound]?"
- "Is [compound] a carcinogen (IARC classification)?"
- "Which genes does [chemical] interact with (CTD)?"
- "Regulatory risk assessment for [environmental chemical]"
- "What diseases are associated with [chemical] exposure?"

Do NOT use for FDA-approved drugs with FAERS data — use `tooluniverse-toxicology` instead.

## Key Tools

| Tool | Purpose | Key Params |
|------|---------|-----------|
| `AOPWiki_list_aops` | Discover AOPs by keyword | `keyword` (organ, effect, or target name) |
| `AOPWiki_get_aop` | Full AOP details: MIE, key events, stressors | `aop_id` (int) |
| `PubChemTox_get_toxicity_summary` | Narrative toxicity overview | `cid` (PubChem CID) |
| `PubChemTox_get_ghs_classification` | GHS hazard category + pictograms | `cid` |
| `PubChemTox_get_carcinogen_classification` | IARC/NTP/EPA carcinogen status | `cid` |
| `PubChemTox_get_toxicity_values` | LD50/LC50 by route and species | `cid` |
| `PubChemTox_get_acute_effects` | Signs and symptoms of acute exposure | `cid` |
| `CTD_get_chemical_gene_interactions` | Chemical-gene molecular interactions | `input_terms` (name or MeSH ID) |
| `CTD_get_chemical_diseases` | Chemical-disease associations | `input_terms` |
| `PubChem_get_CID_by_compound_name` | Resolve compound name to PubChem CID | `name` |

## Workflow

### Phase 1: Compound Identity Resolution

Resolve chemical name to PubChem CID before all PubChemTox calls.

```
PubChem_get_CID_by_compound_name(name="benzo[a]pyrene")
-> cid: 9153 (use for all PubChemTox calls)
```

Note: CTD tools accept the chemical name directly (`input_terms` param) — no CID needed.

### Phase 2: AOP Discovery

Find relevant AOPs by searching organ targets and mechanism keywords.

```
AOPWiki_list_aops(keyword="lung")          # organ-level
AOPWiki_list_aops(keyword="DNA damage")    # mechanism-level
AOPWiki_list_aops(keyword="AhR")           # receptor-level
```

Select 2-4 candidate AOPs from results, then retrieve full details:

```
AOPWiki_get_aop(aop_id=58)  # returns MIE, key events, stressors, biological plausibility
```

Key fields in `AOPWiki_get_aop` response:
- `stressors`: list of chemicals that trigger this AOP (check if query compound is listed)
- `molecular_initiating_event`: the first molecular perturbation
- `key_events`: ordered chain of biological events
- `adverse_outcome`: apical regulatory endpoint

### Phase 3: Hazard Quantification (PubChemTox)

Run all four hazard queries in parallel using the resolved CID:

```
PubChemTox_get_ghs_classification(cid=9153)        # GHS category + pictogram
PubChemTox_get_carcinogen_classification(cid=9153)  # IARC Group 1/2A/2B/3
PubChemTox_get_toxicity_values(cid=9153)            # LD50 by route/species
PubChemTox_get_acute_effects(cid=9153)              # signs/symptoms
```

Note: `PubChemTox_get_target_organs` sometimes returns no data — treat as optional.

### Phase 4: Toxicogenomics (CTD)

Map chemical to gene targets and disease associations:

```
CTD_get_chemical_gene_interactions(input_terms="benzo[a]pyrene")
CTD_get_chemical_diseases(input_terms="benzo[a]pyrene")
```

Cross-reference CTD gene targets with AOP key event genes from Phase 2.

## Tool Parameter Reference

| Tool | Required | Optional | Notes |
|------|---------|---------|-------|
| `AOPWiki_list_aops` | `keyword` | — | Use organ ("liver"), effect ("apoptosis"), or receptor ("PPARalpha") |
| `AOPWiki_get_aop` | `aop_id` | — | Integer ID from list_aops output |
| `PubChemTox_get_toxicity_summary` | `cid` | — | PubChem CID integer |
| `PubChemTox_get_ghs_classification` | `cid` | — | Returns pictogram_labels e.g. "Health Hazard" |
| `PubChemTox_get_carcinogen_classification` | `cid` | — | IARC Group in `classifications[].classification` |
| `PubChemTox_get_toxicity_values` | `cid` | — | Values like "LD50 Rat oral 2400 mg/kg" |
| `PubChemTox_get_acute_effects` | `cid` | — | Sometimes sparse; not all compounds have data |
| `CTD_get_chemical_gene_interactions` | `input_terms` | — | Accepts name or MeSH ID (e.g., "D001564") |
| `CTD_get_chemical_diseases` | `input_terms` | — | Filter `DirectEvidence` = "marker/mechanism" for curated |
| `PubChem_get_CID_by_compound_name` | `name` | — | Returns CID + SMILES; required before PubChemTox calls |

## Common Patterns

```python
# Pattern: Confirm compound is a stressor in a specific AOP
aop = AOPWiki_get_aop(aop_id=58)
stressors = [s["name"] for s in aop["data"]["stressors"]]
# Check if query chemical appears in stressors list

# Pattern: Extract curated CTD disease associations only
diseases = CTD_get_chemical_diseases(input_terms="rotenone")
curated = [d for d in diseases["data"] if d.get("DirectEvidence")]

# Pattern: GHS carcinogen check
carcinogen = PubChemTox_get_carcinogen_classification(cid=9153)
iarc = [c for c in carcinogen["data"]["classifications"] if "IARC" in c.get("source", "")]
```

## Reasoning Framework for Result Interpretation

### Evidence Grading

| Grade | Criteria | Example |
|-------|----------|---------|
| **Strong** | AOP in OECD-endorsed status, compound listed as stressor, CTD + AOPWiki concordant | AOP 58 (AhR → liver tumor) endorsed, benzo[a]pyrene confirmed stressor |
| **Moderate** | AOP under review or well-documented, compound class match but not individually listed | AOP links PPARalpha activation to liver effects; query compound is a fibrate analog |
| **Weak** | AOP in development, compound not listed but shares MIE target via CTD gene overlap | CTD shows gene target overlap with AOP key event genes, but no direct stressor listing |
| **Insufficient** | No AOP found, no CTD gene-disease link, hazard data sparse | Novel compound with no toxicological database entries |

### Interpretation Guidance

- **AOP weight-of-evidence assessment**: OECD-endorsed AOPs have undergone expert review and represent the highest confidence mechanistic pathways. AOPs "under development" in AOPWiki may have incomplete key event relationships. Evaluate each AOP by: (1) biological plausibility of key event relationships, (2) empirical support (dose-response concordance), (3) essentiality of key events (blocking KE prevents AO).
- **Key event relationship (KER) strength**: Strong KERs have dose-response and temporal concordance between upstream and downstream key events. Moderate KERs have correlative evidence. Weak KERs are based on plausibility alone. The weakest KER in the chain determines the overall AOP confidence for that pathway.
- **Stressor potency interpretation**: LD50 values indicate acute toxicity (lower = more toxic). GHS categories: Cat 1 (LD50 <= 5 mg/kg, fatal), Cat 2 (5-50, fatal), Cat 3 (50-300, toxic), Cat 4 (300-2000, harmful), Cat 5 (2000-5000, may be harmful). IARC Group 1 = confirmed carcinogen, 2A = probable, 2B = possible, 3 = not classifiable. Always report route of exposure and species for LD50 values.
- **CTD integration**: CTD "direct evidence" (curated marker/mechanism) is stronger than "inferred" associations. When CTD gene targets overlap with AOP key event genes, this supports the mechanistic link between the compound and the adverse outcome.
- **Regulatory context**: For risk assessment, combine hazard identification (IARC, GHS) with exposure assessment. A potent carcinogen at negligible exposure may pose lower risk than a moderate toxicant at high exposure.

### Synthesis Questions

1. Is the query compound explicitly listed as a stressor in the identified AOP, or is the link inferred from shared molecular targets (CTD gene overlap)?
2. Do the key event relationships in the AOP chain have sufficient empirical support (dose-response concordance, temporal sequence), or are there weak links that reduce confidence?
3. Are the hazard data (LD50, GHS, IARC) consistent across sources, and do they support the severity implied by the AOP adverse outcome?
4. Does the CTD gene-disease evidence corroborate the AOP's predicted adverse outcome, or are there discrepancies suggesting alternative pathways?
5. For regulatory decision-making, is the combined weight of evidence (AOP mechanism + hazard quantification + exposure context) sufficient to support a risk classification?

---

## Fallback Chains

| Primary | Fallback | When |
|---------|---------|------|
| `AOPWiki_list_aops` with specific keyword | Broader organ term | No results |
| `PubChemTox_get_target_organs` | `PubChemTox_get_toxicity_summary` | Returns empty |
| `CTD_get_chemical_diseases` | `CTD_get_gene_diseases` + gene from CTD interactions | Compound name not recognized |

© mims-harvard · Apache-2.0 · mims-harvard/ToolUniverse

Similar skills

All skills for Science and research →

From the same set136

The whole set →
tooluniverse-lipidomics tooluniverse-lipidomics is a Claude Code skill that identifies and classifies lipids, maps metabolic pathways, and retrieves disease associations for lipidomics analysis. CX Codex +1 238installs tooluniverse-comparative-genomics tooluniverse-comparative-genomics is a Claude Code skill that finds orthologs and assesses gene conservation across species. CX Codex +1 238installs tooluniverse-metabolomics-pathway The tooluniverse-metabolomics-pathway skill connects metabolites with pathways, enzymes, genes, and diseases. Use it when metabolomics work needs cross-database identification and mapping. CX Codex +1 238installs tooluniverse-electron-microscopy The tooluniverse-electron-microscopy skill for Claude Code searches and inspects electron microscopy records in EMDB, PDB, EMPIAR, and CryoET Data Portal. It helps link density maps to fitted atomic models. CX Codex +1 238installs tooluniverse-model-organism-genetics tooluniverse-model-organism-genetics is a Claude Code skill for comparing human genes with model organisms, mapping orthologs and retrieving phenotype and expression data. CX Codex +1 235installs tooluniverse-kegg-disease-drug tooluniverse-kegg-disease-drug is a Claude Code skill that links diseases to genes, drugs to targets and variants to pathways via KEGG. It activates on keywords like KEGG disease, drug targets or variant networks. CX Codex +1 235installs tooluniverse-population-genetics-1000genomes tooluniverse-population-genetics-1000genomes is a Claude Code skill for analyzing genetic diversity through 1000 Genomes and comparing variant frequencies across ancestry groups. CX Codex +1 235installs tooluniverse-cancer-classification tooluniverse-cancer-classification maps free-text tumor descriptions to OncoTree codes, links NCI and UMLS IDs, and feeds verified tumor types into OncoKB variant annotations. CX Codex +1 234installs

Project sponsors

Companies that keep this analytics open