Conciseness

The body is dense with RELION-specific operational constraints Claude does not inherently know (one-based 'index@stack.mrcs' indices, project-relative path resolution, odd MPI ranks >= 3, the <=256^3 FSC limit, B-factor left at zero), with no padding or generic-library explanation. A few minor instances could be trimmed — 'Nyquist resolution is 2 × pixel size' and some generic inspection advice — placing it at anchor 4 rather than anchor 5's 'every token earns its place'.

Actionability

Four fully executable, copy-paste-ready commands ('python scripts/spa_workflow.py validate-star ...', 'refine --star ... --mpi-ranks 3', 'fsc ... --output diagnostic-fsc.tsv', 'postprocess --half1 ... --mask ...') cover all subcommands with concrete flags and realistic paths. Illustrative values are explicitly justified ('The diameter and low-pass filter above are illustrative Å values. Use specimen-appropriate values').

Workflow Clarity

Sections follow the actual pipeline order (validate STAR → refine → inspect half maps → postprocess) matching the intro's stated sequence, with pre-flight checklists ('Before running, inspect representative particles and class averages...'), explicit validation semantics ('stops on process failure', 'requires converged unfiltered half maps before reporting success', "'null' means no downward crossing was detected"), a feedback loop ('residual correlation calls for a smoother/wider mask and another postprocessing run'), and restart guidance (use the original job's optimiser).

Progressive Disclosure

A lean ~120-line overview with two clearly signaled, purpose-scoped references ('Read [references/acquisition-and-restarts.md] when starting from movies or resuming jobs'; 'See [references/runtime-and-validation.md] for the tested native utilities and the distinction between pipeline execution and reconstruction validation'), both verified to exist and to contain no further nested references. Detail is appropriately split out, and external primary references are listed at the end.

Specificity

The description lists multiple concrete actions — 'Validates and executes RELION single-particle cryo-EM refinement and half-map postprocessing. Supports STAR optics/acquisition checks, particle-stack consistency, gold-standard half sets, soft-mask validation, diagnostic Fourier shell correlation, and restart guidance' — covering every bundled subcommand (validate, refine, FSC, postprocess) plus restarts. This is comprehensive coverage of the skill's capabilities, exceeding anchor 4's 'several specific actions; minor gaps'.

Completeness

The 'what' is clear and specific, but there is no 'Use when...' clause or equivalent explicit trigger guidance anywhere in the description; the rubric explicitly caps completeness at 3 in that case. Anchor 4 ('when' present but could be more specific) does not fit because no 'when' phrasing exists at all.

Trigger Term Quality

Good natural keyword coverage — 'RELION', 'single-particle', 'cryo-EM', 'refinement', 'postprocessing', 'Fourier shell correlation', 'soft-mask', 'gold-standard' — which a cryo-EM user would naturally say. Anchor 5 requires synonyms and file extensions, but no extensions (.star, .mrc, .mrcs) or common variations like 'map sharpening' appear.

Distinctiveness Conflict Risk

The description occupies a clear niche — RELION single-particle cryo-EM with distinctive terms like 'gold-standard half sets' and 'diagnostic Fourier shell correlation' — making it highly distinguishable from any other skill with minimal conflict risk.