Conciseness

The body is dense and mostly non-redundant domain knowledge (DPP4 liability, interface divergence, non-canonical residue flags), but it repeats material: the exendin-4 validation result is stated in full twice (script section and Validation section), and "A name-level guess ('it's a GLP-1 analog so it's GLP1R') is exactly what produces off-target errors" appears verbatim twice. The ~19-line 'Peptide Bioactivity Property Databases' section is admitted to be peripheral ('None of them tell you what a peptide's target is') and could be trimmed or offloaded. This fits anchor 3 ('includes some unnecessary explanation or could be tightened') better than 4, where over-explanation would be only minor.

Actionability

Two copy-paste-ready script invocations with every flag documented inline ("--sequence <PEPTIDE_SEQ> ... --hypothesized-target <GENE> ... --phenotype '<disease name>'"), concrete mode examples ("--hypothesized-target GLP1R", "--assay-species mus_musculus"), and a fully validated control run with the exact input sequence and its checkable expected output (class-B panel, GIPR Tier 1 score 0.674). This matches anchor 5 — executable commands covering the common cases with specific examples.

Workflow Clarity

The six phases are clearly sequenced in a table with per-phase tools and a core intersection rule, and checkpoints exist: Phase 0 'Verify: confirm every tool loads; substitute fallbacks', a DRY RUN gate before key-gated GPU spend, and fallback behavior ('degrades to phenotype-only if that resolver is transiently down'). It falls short of anchor 5 because the stated verification hook — 'verify the output against that eval's assertions' in evals/evals.json — points to a file that is absent from the bundle, leaving that checkpoint non-executable, and the batch mode gets no per-record validation guidance.

Progressive Disclosure

Good structure: the body is an overview with scripts as the fast path, and the deep material is a single, clearly signaled, one-level reference (references/phases.md, present in the bundle, with explicit when-to-read guidance: 'Read it when a script step fails, when you want to drive a phase manually, or when you extend the pipeline'). The gaps keeping it from anchor 5 are the dangling evals/evals.json reference and the inline 7-database supplementary table (with verified-live-behavior notes) that would sit more naturally in its own reference file — minor organization gaps consistent with anchor 4.

Specificity

The description lists multiple concrete, specific actions — "PROSITE/ELM motif, BLAST homology, HGNC/InterPro/GPCRdb/GtoPdb target-family enumeration, OpenTargets phenotype anchor, EnsemblCompara/Alliance cross-species reconciliation" plus "co-folding (Boltz2, AlphaFold2-Multimer, OpenFold3) for structural confirmation" — covering the full pipeline comprehensively with named tools. It clearly matches the anchor 'Lists multiple specific concrete actions; comprehensive coverage' rather than 4, which requires 'minor gaps in coverage'.

Completeness

It explicitly answers both questions with concrete triggers: what — "Find the real protein target(s) of a peptide from its sequence"; when — "Use when a peptide has a phenotype but does not bind its hypothesized target, when a peptide binds a target in one species or assay but not another, or to screen candidate targets for an orphan peptide". This mirrors anchor 5 exactly; anchor 4 ('when' could be more explicit) is clearly surpassed by the three concrete trigger scenarios.

Trigger Term Quality

Natural user phrasings are comprehensively covered as synonyms: "peptide target deorphanization / off-target identification", "orphan peptide", "screen candidate targets", "does not bind its hypothesized target", "binds a target in one species or assay but not another", plus target-class nouns (GPCR, ion channel, protease, integrin). This matches anchor 5's comprehensive synonym coverage; anchor 4 would require visibly missing common phrasings, and none are apparent for this niche.

Distinctiveness Conflict Risk

Peptide target deorphanization is a clear niche with distinct triggers (orphan peptide, off-target identification, cross-species binding mismatch) that would rarely collide with other skills' triggers. The description also explicitly scopes the boundary ("for ANY target class ... not only GPCRs"), matching anchor 5's 'Clear niche with distinct triggers; minimal conflict risk'.