[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M6DAKX9XF3VGB84J79WRAYD8":3,"skill-same:01M6DAKX9XF3VGB84J79WRAYD8":61},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":18,"publishedAt":30,"linkOnly":31,"markdownContent":32,"markdownBody":33,"frontMatter":34,"refName":35,"commitSha":36,"skillDirectory":37,"forksCount":14,"repositoryHtmlUrl":38,"repositoryHomepageUrl":6,"ownerHtmlUrl":39,"ownerBlogUrl":6,"ownerDescription":6,"updatedAt":40,"install":41,"resources":54},"01M6DAKX9XF3VGB84J79WRAYD8","tooluniverse-admet-prediction",null,"Comprehensive ADMET (Absorption, Distribution, Metabolism, Excretion, Toxicity) profiling of drug candidates using ADMETAI predictions, SwissADME drug-likeness, PubChemTox experimental toxicity, ChEMBL clinical data, and PubChem properties. Generates a structured ADMET scorecard with pass\u002Ffail verdicts per category. Use when asked about drug-likeness, ADMET properties, bioavailability, toxicity prediction, BBB penetration, CYP interactions, pharmacokinetic profiling, Lipinski rule of five, or ADME\u002FPK assessment of a compound.","Prekzursil\u002Fagent-skills-toolchain","Prekzursil","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F54636077?v=4","01M6DACZ1PCQHD99SCR1YXG28B","Andrei Mihai Visalon","01M6DACZ670WZF535V3WYYSS4Z",0,"MIT","https:\u002F\u002Fgithub.com\u002FPrekzursil\u002Fagent-skills-toolchain\u002Fblob\u002Fmain\u002F.agents\u002Flibrary\u002Fskills\u002Ftooluniverse-admet-prediction\u002FSKILL.md",[],[19,20,21,22,23,24,25,26,27,28,29],"tooluniverse admet prediction","assessment","data","profiling","use","properties","compound","structured","metabolism","distribution","clinical","2026-11-10 10:12:30",false,"---\nname: tooluniverse-admet-prediction\ndescription: Comprehensive ADMET (Absorption, Distribution, Metabolism, Excretion, Toxicity) profiling of drug candidates using ADMETAI predictions, SwissADME drug-likeness, PubChemTox experimental toxicity, ChEMBL clinical data, and PubChem properties. Generates a structured ADMET scorecard with pass\u002Ffail verdicts per category. Use when asked about drug-likeness, ADMET properties, bioavailability, toxicity prediction, BBB penetration, CYP interactions, pharmacokinetic profiling, Lipinski rule of five, or ADME\u002FPK assessment of a compound.\n---\n\n# ADMET Prediction & Drug Candidate Profiling\n\n**ADMET reasoning**: a drug fails if it can't be absorbed, distributes to wrong tissues, isn't metabolized safely, or isn't excreted. Evaluate each property independently — good absorption doesn't compensate for liver toxicity. The ADME properties determine whether a compound reaches its target at therapeutic concentrations; toxicity determines whether it's safe to do so. Prioritize experimental data (T2) over computational predictions (T3) — ADMETAI predictions are screening tools, not definitive verdicts. When a FAIL is flagged in any toxicity category (hERG, AMES, DILI), treat it as program-limiting until wet-lab data refutes it.\n\n**LOOK UP DON'T GUESS**: never assume SMILES, CID, or experimental LD50 values — always call PubChem to resolve compound identity before any ADMETAI or PubChemTox call.\n\nComprehensive pharmacokinetic and toxicity profiling integrating AI-based ADMET predictions, rule-based drug-likeness filters, and experimental benchmarks from curated databases.\n\n## When to Use This Skill\n\n**Triggers**:\n- \"What are the ADMET properties of [compound]?\"\n- \"Is [drug] likely to cross the blood-brain barrier?\"\n- \"Predict the toxicity of this SMILES: ...\"\n- \"Does [compound] violate Lipinski's rule of five?\"\n- \"Assess the drug-likeness of [molecule]\"\n- \"What are the CYP interactions for [drug]?\"\n- \"Pharmacokinetic profile of [compound]\"\n- \"Is [compound] orally bioavailable?\"\n- \"What is the LD50 \u002F hERG liability of [molecule]?\"\n\n**Input**: Drug name (e.g., \"ibuprofen\") OR SMILES string (e.g., \"CC(C)Cc1ccc(cc1)C(C)C(=O)O\")\n\n---\n\n## COMPUTE, DON'T DESCRIBE\nWhen analysis requires computation (statistics, data processing, scoring, enrichment), write and run Python code via Bash. Don't describe what you would do — execute it and report actual results. Use ToolUniverse tools to retrieve data, then Python (pandas, scipy, statsmodels, matplotlib) to analyze it.\n\n## KEY PRINCIPLES\n\n1. **Resolve identity first** - Always convert drug name to SMILES before calling ADMETAI tools\n2. **ADMETAI tools require `tooluniverse[ml]`** - If import fails, skip to SwissADME\u002FPubChemTox fallbacks\n3. **All ADMETAI tools take `smiles: list[str]`** - Always wrap in a list, even for one compound\n4. **SwissADME takes `smiles: str`** - Single string, NOT a list (SOAP-style with `operation` param)\n5. **PubChemTox tools accept `cid` or `compound_name`** - Use CID when available for reliability\n6. **Evidence grading mandatory** - Predictions (T3), experimental data (T2), regulatory (T1)\n7. **Scorecard output** - Every analysis must end with a pass\u002Fwarn\u002Ffail scorecard\n8. **Explain significance** - State WHY each property matters for drug development\n\n---\n\n## Evidence Grading\n\n| Tier | Label | Source |\n|------|-------|--------|\n| **T1** | Regulatory\u002FClinical | FDA labels, ChEMBL max clinical phase |\n| **T2** | Experimental | PubChemTox LD50\u002FLC50, in vitro AMES, animal studies |\n| **T3** | Computational | ADMETAI predictions, SwissADME calculations |\n| **T4** | Annotation | Database cross-references, text-mined |\n\n## Workflow: 5-Phase ADMET Profiling\n\n```\nUser Query (drug name or SMILES)\n|\n+-- PHASE 1: Compound Identity Resolution\n|   PubChem name->CID->SMILES, or validate input SMILES\n|\n+-- PHASE 2: Physicochemical & Drug-Likeness\n|   ADMETAI physicochemical + SwissADME druglikeness -> Lipinski\u002FVeber\n|\n+-- PHASE 3: ADME Predictions\n|   BBB, bioavailability, CYP interactions, clearance, solubility\n|\n+-- PHASE 4: Toxicity Assessment\n|   ADMETAI tox + PubChemTox experimental + nuclear receptor + stress\n|\n+-- PHASE 5: Scorecard & Clinical Context\n|   ChEMBL max phase, aggregate pass\u002Fwarn\u002Ffail, final recommendation\n```\n\n---\n\n### PHASE 1: Compound Identity Resolution\n\n**Goal**: Obtain SMILES, PubChem CID, and basic identifiers for the query compound.\n\n**Steps**:\n\n1. **If input is a drug name**:\n   - Call `PubChem_get_CID_by_compound_name(name=\u003Cdrug_name>)` to get CID\n   - Call `PubChem_get_compound_properties_by_CID(cid=\u003CCID>)` to get SMILES and MW\n   - Extract `ConnectivitySMILES` from the response (NOT `CanonicalSMILES`)\n\n2. **If input is a SMILES string**:\n   - Call `PubChem_get_CID_by_SMILES(smiles=\u003CSMILES>)` to get CID\n   - Call `PubChem_get_compound_properties_by_CID(cid=\u003CCID>)` for compound name and MW\n   - Use the input SMILES for all subsequent ADMETAI calls\n\n3. **Record**:\n   - Compound name, CID, SMILES, molecular formula, molecular weight, IUPAC name\n   - If CID lookup fails, proceed with SMILES only (ADMETAI does not need CID)\n\n**Why this matters**: ADMETAI tools require SMILES input. PubChemTox tools work best with CID. Resolving both ensures all downstream tools can be called. PubChem is the authoritative source for SMILES canonicalization.\n\n**Fallback**: If PubChem has no entry, the user must provide SMILES directly. Cannot proceed without SMILES.\n\n---\n\n### PHASE 2: Physicochemical Properties & Drug-Likeness\n\n**Goal**: Evaluate whether the compound has drug-like physicochemical properties.\n\n**Steps**:\n\n1. **ADMETAI physicochemical** (primary):\n   ```\n   ADMETAI_predict_physicochemical_properties(smiles=[\"\u003CSMILES>\"])\n   ```\n   Returns: MW, logP, TPSA, HBD, HBA, rotatable bonds\n\n2. **SwissADME drug-likeness** (complementary):\n   ```\n   SwissADME_check_druglikeness(operation=\"check_druglikeness\", smiles=\"\u003CSMILES>\")\n   SwissADME_calculate_adme(operation=\"calculate_adme\", smiles=\"\u003CSMILES>\")\n   ```\n   Returns: Lipinski, Veber, Ghose, Egan, Muegge rule compliance; PAINS alerts; Brenk alerts\n\n3. **ADMETAI solubility**:\n   ```\n   ADMETAI_predict_solubility_lipophilicity_hydration(smiles=[\"\u003CSMILES>\"])\n   ```\n   Returns: Aqueous solubility (LogS), lipophilicity, hydration free energy\n\n**Interpret & Score**:\n\n| Property | Ideal Range | Why It Matters |\n|----------|-------------|----------------|\n| MW | \u003C 500 Da | Larger molecules have poor membrane permeability (Lipinski) |\n| LogP | -0.4 to 5.6 | Too hydrophobic = poor solubility; too hydrophilic = poor permeability |\n| HBD | \u003C= 5 | Excess donors reduce membrane crossing (Lipinski) |\n| HBA | \u003C= 10 | Excess acceptors reduce membrane crossing (Lipinski) |\n| TPSA | \u003C 140 A^2 | High PSA correlates with poor oral absorption |\n| Rotatable bonds | \u003C= 10 | Molecular flexibility affects bioavailability (Veber) |\n| LogS | > -6 | Below -6 = practically insoluble, formulation challenge |\n| PAINS alerts | 0 | Pan-assay interference compounds give false positives in screens |\n\n**Verdict**: PASS if Lipinski \u003C= 1 violation and no PAINS alerts; WARN if 2 violations; FAIL if 3+ violations or PAINS+.\n\n**Fallback**: If ADMETAI import fails (missing `tooluniverse[ml]`), rely on SwissADME alone. SwissADME provides all Lipinski descriptors independently.\n\n---\n\n### PHASE 3: ADME Predictions\n\n**Goal**: Predict absorption, distribution, metabolism, and excretion behavior.\n\n**Steps**:\n\n1. **Blood-brain barrier penetration**:\n   ```\n   ADMETAI_predict_BBB_penetrance(smiles=[\"\u003CSMILES>\"])\n   ```\n   - BBB+ = compound can cross; BBB- = cannot\n   - Critical for CNS drugs (must cross) and peripherally-acting drugs (should NOT cross to avoid CNS side effects)\n\n2. **Oral bioavailability**:\n   ```\n   ADMETAI_predict_bioavailability(smiles=[\"\u003CSMILES>\"])\n   ```\n   - F20% = at least 20% oral bioavailability; F30% = at least 30%\n   - Low bioavailability means the drug is extensively metabolized or poorly absorbed\n   - F \u003C 20% generally requires non-oral routes (IV, inhaled, topical)\n\n3. **CYP450 interactions**:\n   ```\n   ADMETAI_predict_CYP_interactions(smiles=[\"\u003CSMILES>\"])\n   ```\n   - Reports substrate\u002Finhibitor status for CYP1A2, 2C9, 2C19, 2D6, 3A4\n   - **Why CYP matters**: ~75% of drugs are metabolized by CYP enzymes. Inhibiting CYP3A4 (which metabolizes ~50% of drugs) causes dangerous drug-drug interactions (DDIs). CYP2D6 polymorphisms affect ~25% of drugs -- poor metabolizers accumulate toxic levels\n   - Substrate of CYP2D6 = pharmacogenomic risk (poor\u002Fultra-rapid metabolizers)\n   - Inhibitor of CYP3A4 = high DDI risk (co-administered drugs accumulate)\n\n4. **Clearance and distribution**:\n   ```\n   ADMETAI_predict_clearance_distribution(smiles=[\"\u003CSMILES>\"])\n   ```\n   - VDss (volume of distribution): low (\u003C0.7 L\u002Fkg) = confined to plasma; high (>1 L\u002Fkg) = distributed to tissues\n   - Clearance: high clearance = short half-life, frequent dosing needed\n   - Plasma protein binding (PPB): >95% bound = narrow therapeutic window, DDI risk from displacement\n\n5. **SwissADME pharmacokinetics** (cross-validation):\n   - GI absorption (high\u002Flow), P-gp substrate status, skin permeation (logKp)\n\n**Key flags**: BBB+ for non-CNS drug (WARN: CNS side effects); BBB- for CNS drug (FAIL: won't reach target); F \u003C 20% (WARN: poor oral bioavailability); CYP3A4 inhibitor (WARN: high DDI); CYP2D6 substrate (WARN: pharmacogenomic variability); PPB > 99% (WARN: narrow window); high clearance + low bioavailability (FAIL).\n\n**Fallback**: If ADMETAI unavailable, SwissADME provides GI absorption, BBB permeation (yes\u002Fno), P-gp substrate, and CYP inhibition predictions.\n\n---\n\n### PHASE 4: Toxicity Assessment\n\n**Goal**: Evaluate safety liabilities from both predicted and experimental sources.\n\n**Steps**:\n\n1. **ADMETAI toxicity predictions** [T3]:\n   ```\n   ADMETAI_predict_toxicity(smiles=[\"\u003CSMILES>\"])\n   ```\n   Key endpoints:\n   - **AMES**: Mutagenicity (bacterial reverse mutation test). Positive = potential carcinogen; regulatory agencies require AMES testing for all new drugs\n   - **DILI**: Drug-induced liver injury risk. Leading cause of drug withdrawal (e.g., troglitazone). Positive = hepatotoxicity concern requiring liver function monitoring\n   - **hERG**: hERG potassium channel inhibition. Causes QT prolongation and fatal cardiac arrhythmia. hERG+ = cardiotoxicity liability; multiple drugs withdrawn for this (e.g., terfenadine, cisapride)\n   - **ClinTox**: Clinical trial toxicity \u002F FDA withdrawal risk. Trained on drugs that failed trials or were withdrawn for toxicity\n   - **LD50_Zhu**: Predicted lethal dose (mg\u002Fkg, rat oral). Lower = more acutely toxic\n   - **Skin_Reaction**: Dermal sensitization potential. Important for topical drugs\n   - **Carcinogens_Lagunin**: Carcinogenicity prediction\n\n2. **Nuclear receptor activity** [T3]:\n   ```\n   ADMETAI_predict_nuclear_receptor_activity(smiles=[\"\u003CSMILES>\"])\n   ```\n   - AR (androgen receptor), ER (estrogen receptor), AhR, PPAR-gamma activity\n   - Positive = potential endocrine disruption; critical for chronic-use drugs and environmental chemicals\n\n3. **Stress response pathways** [T3]:\n   ```\n   ADMETAI_predict_stress_response(smiles=[\"\u003CSMILES>\"])\n   ```\n   - p53 activation = DNA damage response (genotoxicity signal)\n   - MMP disruption = mitochondrial toxicity\n   - ATAD5 = DNA repair stress\n   - HSE = heat shock \u002F protein misfolding stress\n\n4. **PubChemTox experimental data** [T2] (call all in parallel):\n   ```\n   PubChemTox_get_toxicity_values(cid=\u003CCID>)\n   PubChemTox_get_ghs_classification(cid=\u003CCID>)\n   PubChemTox_get_acute_effects(cid=\u003CCID>)\n   PubChemTox_get_carcinogen_classification(cid=\u003CCID>)\n   PubChemTox_get_target_organs(cid=\u003CCID>)\n   PubChemTox_get_toxicity_summary(cid=\u003CCID>)\n   ```\n   - Real animal study data (LD50, LC50, NOAEL) anchors computational predictions\n   - GHS classification provides internationally harmonized hazard categories\n   - Carcinogen classification from IARC (Group 1\u002F2A\u002F2B), NTP, EPA\n\n**Key flags**: AMES positive (FAIL: mutagenic); DILI positive (WARN: hepatotox); hERG positive (FAIL: cardiac, often program-killing); ClinTox positive (WARN); LD50 \u003C 50 mg\u002Fkg (FAIL: GHS 1-2); LD50 50-300 mg\u002Fkg (WARN: GHS 3); NR-ER\u002FAR active (WARN: endocrine disruption); p53 active (WARN: genotoxicity); IARC Group 1\u002F2A (FAIL: known\u002Fprobable carcinogen).\n\n**Fallback**: If ADMETAI unavailable, PubChemTox provides experimental toxicity data for known compounds. For novel compounds without PubChem entries, flag as \"no experimental toxicity data available -- computational predictions only.\"\n\n---\n\n### PHASE 5: Scorecard Assembly & Clinical Context\n\n**Goal**: Aggregate all findings into a structured ADMET scorecard with pass\u002Fwarn\u002Ffail verdicts.\n\n**Steps**:\n\n1. **ChEMBL clinical status** [T1] (if drug has ChEMBL ID):\n   ```\n   ChEMBL_get_molecule(chembl_id=\"\u003CCHEMBL_ID>\")\n   ```\n   - Max phase: 4 = approved, 3 = Phase III, 2 = Phase II, 1 = Phase I, 0 = preclinical\n   - Ro5 violations from ChEMBL (independent validation of Lipinski)\n   - First approval year, indication class, black box warning flag\n\n2. **Build the ADMET Scorecard**: produce a table with 13 categories (Physicochemical, Solubility, Absorption, Distribution, Metabolism, Excretion, Tox: Mutagenicity\u002FHepatotoxicity\u002FCardiotoxicity\u002FCarcinogenicity\u002FAcute, Endocrine, Clinical Tox), each with PASS\u002FWARN\u002FFAIL verdict and key finding. Include compound identity header and overall verdict. Tag each finding with evidence tier [T1-T3].\n\n3. **Interpretation narrative**: After the scorecard, provide a 3-5 sentence summary:\n   - Highlight the most critical findings (any FAILs or WARNs)\n   - State whether the compound is suitable for oral administration\n   - Note any DDI risks from CYP interactions\n   - Flag pharmacogenomic concerns (CYP2D6 substrate)\n   - Recommend next steps (e.g., \"hERG patch clamp assay recommended to confirm computational prediction\")\n\n---\n\n## Completeness Checklist (MANDATORY before reporting)\n\nBefore delivering the final scorecard, verify:\n\n- [ ] Compound identity resolved (name, CID, SMILES all present or explicitly noted as unavailable)\n- [ ] Physicochemical properties reported with Lipinski verdict\n- [ ] At least one source for each ADME property (ADMETAI or SwissADME)\n- [ ] All 7 ADMETAI toxicity endpoints reported (or marked N\u002FA with reason)\n- [ ] PubChemTox experimental data checked (even if \"no data found\")\n- [ ] Nuclear receptor and stress response checked (or marked N\u002FA)\n- [ ] Evidence tier tagged for every finding\n- [ ] Scorecard table complete with verdicts for all 13 categories\n- [ ] Overall verdict stated\n- [ ] Interpretation narrative provided with actionable next steps\n","# ADMET Prediction & Drug Candidate Profiling\n\n**ADMET reasoning**: a drug fails if it can't be absorbed, distributes to wrong tissues, isn't metabolized safely, or isn't excreted. Evaluate each property independently — good absorption doesn't compensate for liver toxicity. The ADME properties determine whether a compound reaches its target at therapeutic concentrations; toxicity determines whether it's safe to do so. Prioritize experimental data (T2) over computational predictions (T3) — ADMETAI predictions are screening tools, not definitive verdicts. When a FAIL is flagged in any toxicity category (hERG, AMES, DILI), treat it as program-limiting until wet-lab data refutes it.\n\n**LOOK UP DON'T GUESS**: never assume SMILES, CID, or experimental LD50 values — always call PubChem to resolve compound identity before any ADMETAI or PubChemTox call.\n\nComprehensive pharmacokinetic and toxicity profiling integrating AI-based ADMET predictions, rule-based drug-likeness filters, and experimental benchmarks from curated databases.\n\n## When to Use This Skill\n\n**Triggers**:\n- \"What are the ADMET properties of [compound]?\"\n- \"Is [drug] likely to cross the blood-brain barrier?\"\n- \"Predict the toxicity of this SMILES: ...\"\n- \"Does [compound] violate Lipinski's rule of five?\"\n- \"Assess the drug-likeness of [molecule]\"\n- \"What are the CYP interactions for [drug]?\"\n- \"Pharmacokinetic profile of [compound]\"\n- \"Is [compound] orally bioavailable?\"\n- \"What is the LD50 \u002F hERG liability of [molecule]?\"\n\n**Input**: Drug name (e.g., \"ibuprofen\") OR SMILES string (e.g., \"CC(C)Cc1ccc(cc1)C(C)C(=O)O\")\n\n---\n\n## COMPUTE, DON'T DESCRIBE\nWhen analysis requires computation (statistics, data processing, scoring, enrichment), write and run Python code via Bash. Don't describe what you would do — execute it and report actual results. Use ToolUniverse tools to retrieve data, then Python (pandas, scipy, statsmodels, matplotlib) to analyze it.\n\n## KEY PRINCIPLES\n\n1. **Resolve identity first** - Always convert drug name to SMILES before calling ADMETAI tools\n2. **ADMETAI tools require `tooluniverse[ml]`** - If import fails, skip to SwissADME\u002FPubChemTox fallbacks\n3. **All ADMETAI tools take `smiles: list[str]`** - Always wrap in a list, even for one compound\n4. **SwissADME takes `smiles: str`** - Single string, NOT a list (SOAP-style with `operation` param)\n5. **PubChemTox tools accept `cid` or `compound_name`** - Use CID when available for reliability\n6. **Evidence grading mandatory** - Predictions (T3), experimental data (T2), regulatory (T1)\n7. **Scorecard output** - Every analysis must end with a pass\u002Fwarn\u002Ffail scorecard\n8. **Explain significance** - State WHY each property matters for drug development\n\n---\n\n## Evidence Grading\n\n| Tier | Label | Source |\n|------|-------|--------|\n| **T1** | Regulatory\u002FClinical | FDA labels, ChEMBL max clinical phase |\n| **T2** | Experimental | PubChemTox LD50\u002FLC50, in vitro AMES, animal studies |\n| **T3** | Computational | ADMETAI predictions, SwissADME calculations |\n| **T4** | Annotation | Database cross-references, text-mined |\n\n## Workflow: 5-Phase ADMET Profiling\n\n```\nUser Query (drug name or SMILES)\n|\n+-- PHASE 1: Compound Identity Resolution\n|   PubChem name->CID->SMILES, or validate input SMILES\n|\n+-- PHASE 2: Physicochemical & Drug-Likeness\n|   ADMETAI physicochemical + SwissADME druglikeness -> Lipinski\u002FVeber\n|\n+-- PHASE 3: ADME Predictions\n|   BBB, bioavailability, CYP interactions, clearance, solubility\n|\n+-- PHASE 4: Toxicity Assessment\n|   ADMETAI tox + PubChemTox experimental + nuclear receptor + stress\n|\n+-- PHASE 5: Scorecard & Clinical Context\n|   ChEMBL max phase, aggregate pass\u002Fwarn\u002Ffail, final recommendation\n```\n\n---\n\n### PHASE 1: Compound Identity Resolution\n\n**Goal**: Obtain SMILES, PubChem CID, and basic identifiers for the query compound.\n\n**Steps**:\n\n1. **If input is a drug name**:\n   - Call `PubChem_get_CID_by_compound_name(name=\u003Cdrug_name>)` to get CID\n   - Call `PubChem_get_compound_properties_by_CID(cid=\u003CCID>)` to get SMILES and MW\n   - Extract `ConnectivitySMILES` from the response (NOT `CanonicalSMILES`)\n\n2. **If input is a SMILES string**:\n   - Call `PubChem_get_CID_by_SMILES(smiles=\u003CSMILES>)` to get CID\n   - Call `PubChem_get_compound_properties_by_CID(cid=\u003CCID>)` for compound name and MW\n   - Use the input SMILES for all subsequent ADMETAI calls\n\n3. **Record**:\n   - Compound name, CID, SMILES, molecular formula, molecular weight, IUPAC name\n   - If CID lookup fails, proceed with SMILES only (ADMETAI does not need CID)\n\n**Why this matters**: ADMETAI tools require SMILES input. PubChemTox tools work best with CID. Resolving both ensures all downstream tools can be called. PubChem is the authoritative source for SMILES canonicalization.\n\n**Fallback**: If PubChem has no entry, the user must provide SMILES directly. Cannot proceed without SMILES.\n\n---\n\n### PHASE 2: Physicochemical Properties & Drug-Likeness\n\n**Goal**: Evaluate whether the compound has drug-like physicochemical properties.\n\n**Steps**:\n\n1. **ADMETAI physicochemical** (primary):\n   ```\n   ADMETAI_predict_physicochemical_properties(smiles=[\"\u003CSMILES>\"])\n   ```\n   Returns: MW, logP, TPSA, HBD, HBA, rotatable bonds\n\n2. **SwissADME drug-likeness** (complementary):\n   ```\n   SwissADME_check_druglikeness(operation=\"check_druglikeness\", smiles=\"\u003CSMILES>\")\n   SwissADME_calculate_adme(operation=\"calculate_adme\", smiles=\"\u003CSMILES>\")\n   ```\n   Returns: Lipinski, Veber, Ghose, Egan, Muegge rule compliance; PAINS alerts; Brenk alerts\n\n3. **ADMETAI solubility**:\n   ```\n   ADMETAI_predict_solubility_lipophilicity_hydration(smiles=[\"\u003CSMILES>\"])\n   ```\n   Returns: Aqueous solubility (LogS), lipophilicity, hydration free energy\n\n**Interpret & Score**:\n\n| Property | Ideal Range | Why It Matters |\n|----------|-------------|----------------|\n| MW | \u003C 500 Da | Larger molecules have poor membrane permeability (Lipinski) |\n| LogP | -0.4 to 5.6 | Too hydrophobic = poor solubility; too hydrophilic = poor permeability |\n| HBD | \u003C= 5 | Excess donors reduce membrane crossing (Lipinski) |\n| HBA | \u003C= 10 | Excess acceptors reduce membrane crossing (Lipinski) |\n| TPSA | \u003C 140 A^2 | High PSA correlates with poor oral absorption |\n| Rotatable bonds | \u003C= 10 | Molecular flexibility affects bioavailability (Veber) |\n| LogS | > -6 | Below -6 = practically insoluble, formulation challenge |\n| PAINS alerts | 0 | Pan-assay interference compounds give false positives in screens |\n\n**Verdict**: PASS if Lipinski \u003C= 1 violation and no PAINS alerts; WARN if 2 violations; FAIL if 3+ violations or PAINS+.\n\n**Fallback**: If ADMETAI import fails (missing `tooluniverse[ml]`), rely on SwissADME alone. SwissADME provides all Lipinski descriptors independently.\n\n---\n\n### PHASE 3: ADME Predictions\n\n**Goal**: Predict absorption, distribution, metabolism, and excretion behavior.\n\n**Steps**:\n\n1. **Blood-brain barrier penetration**:\n   ```\n   ADMETAI_predict_BBB_penetrance(smiles=[\"\u003CSMILES>\"])\n   ```\n   - BBB+ = compound can cross; BBB- = cannot\n   - Critical for CNS drugs (must cross) and peripherally-acting drugs (should NOT cross to avoid CNS side effects)\n\n2. **Oral bioavailability**:\n   ```\n   ADMETAI_predict_bioavailability(smiles=[\"\u003CSMILES>\"])\n   ```\n   - F20% = at least 20% oral bioavailability; F30% = at least 30%\n   - Low bioavailability means the drug is extensively metabolized or poorly absorbed\n   - F \u003C 20% generally requires non-oral routes (IV, inhaled, topical)\n\n3. **CYP450 interactions**:\n   ```\n   ADMETAI_predict_CYP_interactions(smiles=[\"\u003CSMILES>\"])\n   ```\n   - Reports substrate\u002Finhibitor status for CYP1A2, 2C9, 2C19, 2D6, 3A4\n   - **Why CYP matters**: ~75% of drugs are metabolized by CYP enzymes. Inhibiting CYP3A4 (which metabolizes ~50% of drugs) causes dangerous drug-drug interactions (DDIs). CYP2D6 polymorphisms affect ~25% of drugs -- poor metabolizers accumulate toxic levels\n   - Substrate of CYP2D6 = pharmacogenomic risk (poor\u002Fultra-rapid metabolizers)\n   - Inhibitor of CYP3A4 = high DDI risk (co-administered drugs accumulate)\n\n4. **Clearance and distribution**:\n   ```\n   ADMETAI_predict_clearance_distribution(smiles=[\"\u003CSMILES>\"])\n   ```\n   - VDss (volume of distribution): low (\u003C0.7 L\u002Fkg) = confined to plasma; high (>1 L\u002Fkg) = distributed to tissues\n   - Clearance: high clearance = short half-life, frequent dosing needed\n   - Plasma protein binding (PPB): >95% bound = narrow therapeutic window, DDI risk from displacement\n\n5. **SwissADME pharmacokinetics** (cross-validation):\n   - GI absorption (high\u002Flow), P-gp substrate status, skin permeation (logKp)\n\n**Key flags**: BBB+ for non-CNS drug (WARN: CNS side effects); BBB- for CNS drug (FAIL: won't reach target); F \u003C 20% (WARN: poor oral bioavailability); CYP3A4 inhibitor (WARN: high DDI); CYP2D6 substrate (WARN: pharmacogenomic variability); PPB > 99% (WARN: narrow window); high clearance + low bioavailability (FAIL).\n\n**Fallback**: If ADMETAI unavailable, SwissADME provides GI absorption, BBB permeation (yes\u002Fno), P-gp substrate, and CYP inhibition predictions.\n\n---\n\n### PHASE 4: Toxicity Assessment\n\n**Goal**: Evaluate safety liabilities from both predicted and experimental sources.\n\n**Steps**:\n\n1. **ADMETAI toxicity predictions** [T3]:\n   ```\n   ADMETAI_predict_toxicity(smiles=[\"\u003CSMILES>\"])\n   ```\n   Key endpoints:\n   - **AMES**: Mutagenicity (bacterial reverse mutation test). Positive = potential carcinogen; regulatory agencies require AMES testing for all new drugs\n   - **DILI**: Drug-induced liver injury risk. Leading cause of drug withdrawal (e.g., troglitazone). Positive = hepatotoxicity concern requiring liver function monitoring\n   - **hERG**: hERG potassium channel inhibition. Causes QT prolongation and fatal cardiac arrhythmia. hERG+ = cardiotoxicity liability; multiple drugs withdrawn for this (e.g., terfenadine, cisapride)\n   - **ClinTox**: Clinical trial toxicity \u002F FDA withdrawal risk. Trained on drugs that failed trials or were withdrawn for toxicity\n   - **LD50_Zhu**: Predicted lethal dose (mg\u002Fkg, rat oral). Lower = more acutely toxic\n   - **Skin_Reaction**: Dermal sensitization potential. Important for topical drugs\n   - **Carcinogens_Lagunin**: Carcinogenicity prediction\n\n2. **Nuclear receptor activity** [T3]:\n   ```\n   ADMETAI_predict_nuclear_receptor_activity(smiles=[\"\u003CSMILES>\"])\n   ```\n   - AR (androgen receptor), ER (estrogen receptor), AhR, PPAR-gamma activity\n   - Positive = potential endocrine disruption; critical for chronic-use drugs and environmental chemicals\n\n3. **Stress response pathways** [T3]:\n   ```\n   ADMETAI_predict_stress_response(smiles=[\"\u003CSMILES>\"])\n   ```\n   - p53 activation = DNA damage response (genotoxicity signal)\n   - MMP disruption = mitochondrial toxicity\n   - ATAD5 = DNA repair stress\n   - HSE = heat shock \u002F protein misfolding stress\n\n4. **PubChemTox experimental data** [T2] (call all in parallel):\n   ```\n   PubChemTox_get_toxicity_values(cid=\u003CCID>)\n   PubChemTox_get_ghs_classification(cid=\u003CCID>)\n   PubChemTox_get_acute_effects(cid=\u003CCID>)\n   PubChemTox_get_carcinogen_classification(cid=\u003CCID>)\n   PubChemTox_get_target_organs(cid=\u003CCID>)\n   PubChemTox_get_toxicity_summary(cid=\u003CCID>)\n   ```\n   - Real animal study data (LD50, LC50, NOAEL) anchors computational predictions\n   - GHS classification provides internationally harmonized hazard categories\n   - Carcinogen classification from IARC (Group 1\u002F2A\u002F2B), NTP, EPA\n\n**Key flags**: AMES positive (FAIL: mutagenic); DILI positive (WARN: hepatotox); hERG positive (FAIL: cardiac, often program-killing); ClinTox positive (WARN); LD50 \u003C 50 mg\u002Fkg (FAIL: GHS 1-2); LD50 50-300 mg\u002Fkg (WARN: GHS 3); NR-ER\u002FAR active (WARN: endocrine disruption); p53 active (WARN: genotoxicity); IARC Group 1\u002F2A (FAIL: known\u002Fprobable carcinogen).\n\n**Fallback**: If ADMETAI unavailable, PubChemTox provides experimental toxicity data for known compounds. For novel compounds without PubChem entries, flag as \"no experimental toxicity data available -- computational predictions only.\"\n\n---\n\n### PHASE 5: Scorecard Assembly & Clinical Context\n\n**Goal**: Aggregate all findings into a structured ADMET scorecard with pass\u002Fwarn\u002Ffail verdicts.\n\n**Steps**:\n\n1. **ChEMBL clinical status** [T1] (if drug has ChEMBL ID):\n   ```\n   ChEMBL_get_molecule(chembl_id=\"\u003CCHEMBL_ID>\")\n   ```\n   - Max phase: 4 = approved, 3 = Phase III, 2 = Phase II, 1 = Phase I, 0 = preclinical\n   - Ro5 violations from ChEMBL (independent validation of Lipinski)\n   - First approval year, indication class, black box warning flag\n\n2. **Build the ADMET Scorecard**: produce a table with 13 categories (Physicochemical, Solubility, Absorption, Distribution, Metabolism, Excretion, Tox: Mutagenicity\u002FHepatotoxicity\u002FCardiotoxicity\u002FCarcinogenicity\u002FAcute, Endocrine, Clinical Tox), each with PASS\u002FWARN\u002FFAIL verdict and key finding. Include compound identity header and overall verdict. Tag each finding with evidence tier [T1-T3].\n\n3. **Interpretation narrative**: After the scorecard, provide a 3-5 sentence summary:\n   - Highlight the most critical findings (any FAILs or WARNs)\n   - State whether the compound is suitable for oral administration\n   - Note any DDI risks from CYP interactions\n   - Flag pharmacogenomic concerns (CYP2D6 substrate)\n   - Recommend next steps (e.g., \"hERG patch clamp assay recommended to confirm computational prediction\")\n\n---\n\n## Completeness Checklist (MANDATORY before reporting)\n\nBefore delivering the final scorecard, verify:\n\n- [ ] Compound identity resolved (name, CID, SMILES all present or explicitly noted as unavailable)\n- [ ] Physicochemical properties reported with Lipinski verdict\n- [ ] At least one source for each ADME property (ADMETAI or SwissADME)\n- [ ] All 7 ADMETAI toxicity endpoints reported (or marked N\u002FA with reason)\n- [ ] PubChemTox experimental data checked (even if \"no data found\")\n- [ ] Nuclear receptor and stress response checked (or marked N\u002FA)\n- [ ] Evidence tier tagged for every finding\n- [ ] Scorecard table complete with verdicts for all 13 categories\n- [ ] Overall verdict stated\n- [ ] Interpretation narrative provided with actionable next steps\n",{"name":5,"description":7},"main","a298c929b6cbd662e1fce54302ce520d5ceea1a5",".agents\u002Flibrary\u002Fskills\u002Ftooluniverse-admet-prediction","https:\u002F\u002Fgithub.com\u002FPrekzursil\u002Fagent-skills-toolchain","https:\u002F\u002Fgithub.com\u002FPrekzursil","2026-11-02 14:30:57",{"source":8,"skillName":5,"commands":42,"prompt":52,"zipAvailable":53},[43,46,49],{"tool":44,"command":45},"npx","npx skills add Prekzursil\u002Fagent-skills-toolchain --skill tooluniverse-admet-prediction",{"tool":47,"command":48},"bunx","bunx skills add Prekzursil\u002Fagent-skills-toolchain --skill tooluniverse-admet-prediction",{"tool":50,"command":51},"pnpm","pnpm dlx skills add Prekzursil\u002Fagent-skills-toolchain --skill tooluniverse-admet-prediction","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002FPrekzursil\u002Fagent-skills-toolchain\u002Fblob\u002Fmain\u002F.agents\u002Flibrary\u002Fskills\u002Ftooluniverse-admet-prediction\u002FSKILL.md",true,[55],{"path":56,"resourceType":57,"resourceTypeDesc":58,"sizeBytes":59,"sourceHtmlUrl":60},"customize.toml","OTHER","其他",1621,"https:\u002F\u002Fgithub.com\u002Ftonygomes-wq\u002Fevo\u002Fblob\u002Fmain\u002F.kiro\u002Fskills\u002Fgds-retrospective\u002Fcustomize.toml",[62,83,114,141,155,171],{"skillNo":63,"name":64,"nameZh":6,"description":65,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":66,"categories":67,"keywords":68,"publishedAt":30},"01M6DF7D7478R55P46REYH4QAC","clarify-requirements","Resolve genuine pre-execution ambiguity by scanning repo context, generating 1-6 grounded options, and asking 1-3 focused questions via AskUserQuestion. NOT for mid-task use (injection risk); NOT when the goal is already unambiguous; use to disambiguate vague prompts before any file edits.","https:\u002F\u002Fgithub.com\u002FPrekzursil\u002Fagent-skills-toolchain\u002Fblob\u002Fmain\u002Ftools\u002Fworkspace-hooks\u002Fskills\u002Fclarify-requirements\u002FSKILL.md",[],[69,70,71,72,73,74,75,76,77,23,78,79,80,81,82],"clarify requirements","clarify","requirements","risk","options","task","prompts","goal","scanning","context","repo","injection","execution","file",{"skillNo":84,"name":85,"nameZh":6,"description":86,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":87,"categories":88,"keywords":94,"publishedAt":30},"01M6DF79CA4BHJS66J8EZDBPWJ","openmm-md","Run FREE, open-source, Windows-native GPU molecular dynamics via the openmm-md MCP (OpenMM + CUDA on the local NVIDIA GPU). Use for protein \u002F protein-ligand MD when you want a free stack or are on Windows without a Schrodinger license. Covers the full workflow (prepare, solvate, minimize, equilibrate, production, analyze, mmgbsa) AND an honest decision guide vs Schrodinger Desmond, including how to run Desmond legitimately on Linux or a cloud GPU with a real license. This is a deliberately LESSER free alternative and does NOT replace the schrodinger MCP tools. Do NOT use it to pick a free tool when a validated FEP+, WaterMap, IFD-MD, or Prime MM-GBSA result is actually required.","https:\u002F\u002Fgithub.com\u002FPrekzursil\u002Fagent-skills-toolchain\u002Fblob\u002Fmain\u002Ftools\u002Fmcp\u002Fopenmm-md-mcp\u002Fskill\u002Fopenmm-md\u002FSKILL.md",[89],{"categoryNo":90,"categoryName":91,"taxonomyType":92,"taxonomyTypeDesc":93},"01JBSEEDCAT000000000000023","Skill与智能体开发","SCENARIO","使用场景",[95,96,97,98,99,100,101,102,103,104,105,106,107,108,23,109,110,111,112,113],"openmm md","how","decision","pick","linux","analyze","tool","stack","open","NVIDIA","gpu","license","workflow","windows","free","molecular dynamics","local","do","MCP",{"skillNo":115,"name":116,"nameZh":6,"description":117,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":118,"categories":119,"keywords":126,"publishedAt":30},"01M6DF76C1MG85VEWP8VZCBNMN","visual-review","Take screenshots of web pages and UI using Playwright for visual review and iteration","https:\u002F\u002Fgithub.com\u002FPrekzursil\u002Fagent-skills-toolchain\u002Fblob\u002Fmain\u002Ftools\u002Feditor\u002Fmetaswarm\u002Fskills\u002Fvisual-review\u002FSKILL.md",[120,123],{"categoryNo":121,"categoryName":122,"taxonomyType":92,"taxonomyTypeDesc":93},"01JBSEEDCAT000000000000003","演示文稿",{"categoryNo":124,"categoryName":125,"taxonomyType":92,"taxonomyTypeDesc":93},"01JBSEEDCAT000000000000008","测试与质量",[127,128,129,130,131,132,133,134,135,136,96,137,138,139,140],"review","visual","visual review","pages","web","false","iteration","take","check","slides","ui","screenshot","playwright","screenshots",{"skillNo":142,"name":143,"nameZh":6,"description":144,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":145,"categories":146,"keywords":147,"publishedAt":30},"01M6DF73AQVCBZKEN0YXEX9FXP","status","Diagnostic status report — shows metaswarm installation state, project setup, and potential issues","https:\u002F\u002Fgithub.com\u002FPrekzursil\u002Fagent-skills-toolchain\u002Fblob\u002Fmain\u002Ftools\u002Feditor\u002Fmetaswarm\u002Fskills\u002Fstatus\u002FSKILL.md",[],[143,148,149,150,151,152,153,154],"setup","report","project","installation","project setup","issues","state",{"skillNo":156,"name":157,"nameZh":6,"description":158,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":159,"categories":160,"keywords":162,"publishedAt":30},"01M6DF6X6C2XB0P6GX76P7TTYV","start","Use when starting work on any task, when the user mentions metaswarm, or when the user wants to begin tracked development work","https:\u002F\u002Fgithub.com\u002FPrekzursil\u002Fagent-skills-toolchain\u002Fblob\u002Fmain\u002Ftools\u002Feditor\u002Fmetaswarm\u002Fskills\u002Fstart\u002FSKILL.md",[161],{"categoryNo":90,"categoryName":91,"taxonomyType":92,"taxonomyTypeDesc":93},[157,163,74,164,165,166,23,167,168,169,170],"start issue","issue","ready","work on issue","user","work","work on","development",{"skillNo":172,"name":148,"nameZh":6,"description":173,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":174,"categories":175,"keywords":176,"publishedAt":30},"01M6DF6QMJ9AJKFKHHKRT005PQ","Interactive project setup — detects your project, configures metaswarm, writes project-local files","https:\u002F\u002Fgithub.com\u002FPrekzursil\u002Fagent-skills-toolchain\u002Fblob\u002Fmain\u002Ftools\u002Feditor\u002Fmetaswarm\u002Fskills\u002Fsetup\u002FSKILL.md",[],[148,150,111,177,152,178],"files","interactive"]