develop-gc-method
について
このClaudeスキルは、開発者がガスクロマトグラフィー法を一から作成するのを支援し、カラム選択、温度プログラミング、検出器設定を段階的に案内します。新規分析のセットアップや、既存の分析法を異なる機器やサンプルマトリックスに適応させるために設計されています。揮発性および半揮発性の目標化合物に対する初期性能検証も実施します。
クイックインストール
Claude Code
推奨npx skills add pjt222/agent-almanac -a claude-code/plugin add https://github.com/pjt222/agent-almanacgit clone https://github.com/pjt222/agent-almanac.git ~/.claude/skills/develop-gc-methodこのコマンドをClaude Codeにコピー&ペーストしてスキルをインストールします
ドキュメント
Develop a GC Method
Systematic GC method dev: column pick + temp program + carrier + detector + initial perf check for volatile/semi-volatile analytes.
Use When
- New GC for volatile/semi-volatile compounds
- Adapt published method → different instrument/matrix
- Replace existing method failing perf
- Method for compounds w/ known bp + polarity
- Packed → capillary transition
In
Required
- Target analytes: Compounds + CAS + MW + bp
- Sample matrix: Sample type (air, water extract, solvent, bio fluid)
- Detection limits: LOD/LOQ per analyte
Optional
- Reference method: Published (EPA, ASTM, pharmacopeial) → start
- Available columns: On-hand inventory
- Instrument config: GC model, detectors, autosampler
- Throughput: Max run time/sample
- Regulatory: GLP, GMP, EPA, etc
Do
Step 1: Analytical Objectives
- List analytes + props (bp, polarity, MW).
- ID matrix + expected interferents/co-extractives.
- Specify LOD/LOQ, quant range, Rs for critical pairs.
- Method must meet regulatory (EPA 8260, USP, etc)?
- Doc throughput: max run time, inj vol, prep constraints.
→ Written spec: analytes + matrix + limits + Rs + regulatory/throughput.
If err: volatility data missing → estimate bp from structural analogs or scout run on mid-polarity col for elution order.
Step 2: Pick Column
Dimensions + phase via analyte polarity + separation diff.
| Column Type | Stationary Phase | Polarity | Typical Use Cases |
|---|---|---|---|
| DB-1 / HP-1 | 100% dimethylpolysiloxane | Non-polar | Hydrocarbons, solvents, general screening |
| DB-5 / HP-5 | 5% phenyl-methylpolysiloxane | Low polarity | Semi-volatiles, EPA 8270, drugs of abuse |
| DB-1701 | 14% cyanopropylphenyl | Mid polarity | Pesticides, herbicides |
| DB-WAX / HP-INNOWax | Polyethylene glycol | Polar | Alcohols, fatty acids, flavors, essential oils |
| DB-624 | 6% cyanopropylphenyl | Mid polarity | Volatile organics, EPA 624/8260 |
| DB-FFAP | Modified PEG (nitroterephthalic acid) | Highly polar | Organic acids, free fatty acids |
| DB-35 | 35% phenyl-methylpolysiloxane | Mid-low polarity | Polychlorinated biphenyls, confirmatory column |
- Analyte polarity ↔ phase: like dissolves like.
- Length (15-60 m): longer → more plates, longer runs.
- ID (0.25-0.53 mm): narrower → better eff, wider → more capacity.
- Film (0.25-5.0 um): thicker → retain volatiles longer.
- Complex matrices → guard col or retention gap.
→ Col spec (phase + L + ID + film) justified by analyte + Rs reqs.
If err: no single col resolves all → confirm col w/ orthogonal selectivity (DB-1 primary, DB-WAX confirm).
Step 3: Optimize Temp Program
- Initial oven ≤ bp of most volatile (hold 1-2 min → solvent focus).
- Linear ramp starts:
- Simple: 10-20 C/min
- Complex: 3-8 C/min (better Rs)
- Ultra-fast: 25-40 C/min on short thin-film
- Final temp 10-20 C above bp of least volatile.
- Final hold 2-5 min → full elution + bake-out.
- Co-eluting critical pairs → isothermal hold before elution, or slower ramp there.
- Verify total run time meets throughput.
→ Temp program (init + hold + ramp + final + hold) separates all targets in acceptable time.
If err: critical pairs still unresolved after ramp → revisit col (Step 2) or multi-ramp w/ slower rates in problem region.
Step 4: Pick Carrier Gas
| Property | Helium (He) | Hydrogen (H2) | Nitrogen (N2) |
|---|---|---|---|
| Optimal linear velocity | 20-40 cm/s | 30-60 cm/s | 10-20 cm/s |
| Efficiency at high flow | Good | Best (flat van Deemter) | Poor |
| Speed advantage | Baseline | 1.5-2x faster than He | Slowest |
| Safety | Inert | Flammable (needs leak detection) | Inert |
| Cost / availability | Expensive, supply concerns | Low cost, generator option | Very low cost |
| Detector compatibility | All detectors | Not with ECD; caution with some MS | All detectors |
- Default He for general + regulatory methods specifying He.
- H2 for faster or when He supply constrained; install H2 leak detection + interlocks.
- N2 only for simple separations or when cost primary.
- Flow → optimal linear velocity for gas + col ID.
- Measure actual velocity via unretained (e.g., methane on FID).
→ Carrier picked + flow at optimal velocity, verified by unretained peak.
If err: eff lower than expected → van Deemter curve (plate height vs velocity) over 5-7 flows for true optimum.
Step 5: Pick Detector
| Detector | Selectivity | Sensitivity (approx.) | Linear Range | Best For |
|---|---|---|---|---|
| FID | C-H bonds (universal organic) | Low pg C/s | 10^7 | Hydrocarbons, general organics, quantitation |
| TCD | Universal (all compounds) | Low ng | 10^5 | Permanent gases, bulk analysis |
| ECD | Electronegative groups (halogens, nitro) | Low fg (Cl compounds) | 10^4 | Pesticides, PCBs, halogenated solvents |
| NPD/FPD | N, P (NPD); S, P (FPD) | Low pg | 10^4-10^5 | Organophosphorus pesticides, sulfur compounds |
| MS (EI) | Structural identification | Low pg (scan), fg (SIM) | 10^5-10^6 | Unknowns, confirmation, trace analysis |
| MS/MS | Highest selectivity | fg range | 10^5 | Complex matrices, ultra-trace, forensic |
- Match detector to analyte chem + sensitivity.
- Quant in simple matrices → FID default (robust + linear + low maint).
- Trace in complex matrices → MS SIM or MS/MS MRM.
- Halogenated at trace → ECD best sensitivity.
- Detector temp 20-50 C above max oven → prevent condensation.
- Optimize detector gas flows per mfr.
→ Detector picked + config w/ temps + flows for targets.
If err: sensitivity insufficient → concentrate sample (bigger inj, solvent evap) or more sensitive/selective detector.
Step 6: Validate Initial Perf
- System suitability std: all targets at mid-range conc.
- Inject std 6× consec.
- Evaluate:
- RT RSD: < 1.0%
- Peak area RSD: < 2.0% (< 5.0% trace)
- Rs critical pairs: ≥ 1.5 (baseline) or ≥ 2.0 regulated
- Tailing factor: 0.8-1.5 (USP T ≤ 2.0)
- Theoretical plates N: vs col mfr spec
- Blank inj → no carryover/ghost peaks.
- Matrix blank → ID interferents at target RT.
- Doc all in method summary.
→ Suitability met across replicates, no carryover/matrix interference at target windows.
If err: tailing → check active sites (recondition, trim 0.5 m inlet, replace liner). RSD high → autosampler precision + inj technique. Rs low → Step 3 temp refinement.
Check
- All targets Rs ≥ 1.5 critical pairs
- RT RSD < 1.0% over 6 reps
- Peak area RSD < 2.0% over 6 reps
- Tailing 0.8-1.5 all analytes
- Blank no carryover >0.1% working conc
- Matrix blank no interference at targets
- Run time meets throughput
- All params documented (col, temps, flows, detector)
Traps
- Column bleed temp limit: Above max isothermal → elevated baseline + ghost peaks + col degradation. Check spec sheet.
- Oversized inj: Too much solvent → fronting + poor Rs early. Match inj vol to col capacity (0.5-2 uL for 0.25 mm ID split).
- Wrong liner: Splitless → single/double-taper deactivated; split → w/ glass wool. Mismatch → poor repro.
- Septum/liner maint: Coring + contamination = top sources of ghost peaks + tailing. Septa every 50-100 inj, liners scheduled.
- Skip van Deemter: Mfr default flow not measured optimum → wasted eff, esp carrier gas swaps.
- Insufficient conditioning: New cols → condition (ramp to max temp under carrier, no detector) to clear mfr residues.
→
develop-hplc-method— LC method dev for non-volatile/thermally labileinterpret-chromatogram— reading GC + HPLC chromatogramstroubleshoot-separation— diagnose peak shape/RT/Rs problemsvalidate-analytical-method— formal ICH Q2 valid. of GC method
GitHub リポジトリ
関連スキル
qmd
開発qmdは、BM25、ベクトル埋め込み、およびリランキングを組み合わせたハイブリッド検索を用いて、ローカルファイルのインデックス作成と検索を可能にするローカル検索・インデックス作成CLIツールです。コマンドラインでの使用と、Claudeとの統合のためのMCP(Model Context Protocol)モードの両方をサポートしています。このツールは埋め込みにOllamaを使用し、インデックスをローカルに保存するため、ターミナルから直接ドキュメントやコードベースを検索するのに最適です。
subagent-driven-development
開発このスキルは、各独立したタスクに対して新規のサブエージェントを起動し、タスク間でコードレビューを実施しながら実装計画を実行します。レビュープロセスを通じて品質基準を維持しつつ、迅速な反復を可能にします。同一セッション内で主に独立したタスクに取り組む際に本スキルをご利用いただくことで、組み込まれた品質チェックを伴う継続的な進捗を確保できます。
mcporter
開発mcporterスキルは、開発者がClaudeから直接Model Context Protocol(MCP)サーバーを管理および呼び出せるようにします。このスキルは、利用可能なサーバーの一覧表示、引数を指定したツールの呼び出し、認証およびデーモンのライフサイクル管理を行うコマンドを提供します。開発ワークフローにおいてMCPサーバーの機能を統合およびテストする際に、このスキルをご利用ください。
adk-deployment-specialist
開発このスキルは、A2Aプロトコルを使用してVertex AI ADKエージェントをデプロイおよびオーケストレーションし、AgentCardの発見、タスク送信、およびコード実行サンドボックスやメモリバンクなどのサポートツールを管理します。Python、Java、またはGoで、順次、並列、またはループのオーケストレーションパターンを用いたマルチエージェントシステムの構築を可能にします。Google Cloud上でADKエージェントのデプロイやエージェントワークフローのオーケストレーションを求められた際にご利用ください。
