Analytical Method Validation

What it does

• Fixes the governing framework first: the same assay validates differently under ICH Q2(R2), USP <1225>, ICH M10, CLSI EP and ISO/IEC 17025 — blending them yields a protocol that satisfies none.
• Builds the protocol: plan_validation.py maps the measured attribute (assay, quantitative impurity, limit test, identity) to required characteristics, study layout and a protocol skeleton that forces acceptance criteria to be stated before data collection.
• Runs the right statistics: lack-of-fit F test, residual runs test and back-calculated error instead of r² for response; a one-way random-effects model that separates repeatability from between-day intermediate precision; DL/QL by each allowed σ approach plus a confirmation step.
• ICH M10 bioanalysis: modality-specific run checks for chromatographic vs ligand-binding assays (±15% vs ±20%, total error, ISR), including the per-level ≥50% QC rule that overall pass rates hide.
• Method comparison and transfer: TOST equivalence at a pre-stated margin, Deming and Passing–Bablok regression alongside OLS for contrast, Bland–Altman and proportional-bias flags.
• Python 3.11+ standard library only, no network. Data to stdout, provenance and caveats to stderr, exit codes suitable for gating a workflow.

Who it's for

• Analytical development and QC scientists in pharma/biotech, CMC document authors
• Clinical laboratories following CLSI EP and ISO 17025 accredited labs
• CRO bioanalysts running PK/TK/BE studies
• QA staff preparing regulatory submissions or defending audit findings

Examples

1. "Draft an ICH Q2(R2) validation protocol for an HPLC assay" → required-characteristic table, 80–120% reportable range, and bracketed criteria fields left for you to decide.
2. "My calibration has r² = 0.983 — is it fine?" → flags lack-of-fit p = 1.5e-06 and −9.0% back-calculated error at the low end, and suggests a 1/x² weighted fit.
3. "The LBA run passes 2/3 of QCs overall — accept it?" → reports that the high level is 0/2, failing the per-level 50% requirement.
4. "The paired t-test is p > 0.05, so the methods are equivalent, right?" → replaces that with TOST at a ±2% margin and reports Deming/Passing–Bablok slopes instead of biased OLS.

Boundaries

The scripts report; they never conclude that a procedure is validated, release a batch, or close an investigation. Paywalled USP, CLSI and ISO thresholds are intentionally not reproduced — only designations, scope and where to obtain an authorised copy.

Read the instructions in this file and follow them to help me:
https://raw.githubusercontent.com/K-Dense-AI/scientific-agent-skills/HEAD/skills/analytical-method-validation/SKILL.md

What I want: (describe your task here)

Install the skill I found on Claude Skill Mart.
Copy the skills/analytical-method-validation folder from the GitHub repo K-Dense-AI/scientific-agent-skills into my ~/.claude/skills/analytical-method-validation/.
When it's done, tell me in one line what this skill can do.