Content
85%Weight 40%Scale 1-5Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
A well-structured, tool-specific guide: executable commands with realistic payloads, a design-rules table, literal error-message→fix mappings, and an explicit genome-specificity checkpoint the tools don't provide. The main gaps are the missing invocation example for the QC script and a "Why" column that over-explains knowledge Claude already has.
Suggestions
Add a one-line runnable example for scripts/primer_qc.py (e.g. `python scripts/primer_qc.py FWD_SEQ REV_SEQ`) so the QC path — advertised for "primers from any source" — is as executable as the design path.
Drop or compress the table's "Why" column to property/target values only; the rationale (GC balance, Tm matching) is standard molecular biology Claude already knows.
Show a minimal `tu run IDT_analyze_oligo` example alongside the prose parameter list, since hairpin/dimer screening is one of the skill's advertised capabilities.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is lean and tool-focused, but the table's "Why" column ("balanced stability", "mismatched Tm → one primer anneals poorly", "long enough for specificity, short enough for efficient annealing") restates standard molecular biology Claude already knows, fitting the 4 anchor (efficient with minor over-explanation that could be trimmed). It is not 5 because trimming the Why column to property+target would lose nothing Claude can't supply; it is well above 3 since there is no padded prose or concept introduction. | 4 / 5 |
Actionability | The two main workflows are copy-paste-ready `tu run` commands with realistic parameter values, and tool parameters for IDT/NEB are named inline. It falls short of the 5 anchor because `scripts/primer_qc.py` — presented as the QC mechanism "for primers from any source" — has no invocation example, and `IDT_analyze_oligo` / `DNA_calculate_gc_content` are described but never shown as commands. | 4 / 5 |
Workflow Clarity | Steps 1–5 are clearly sequenced (design → Tm/Ta → QC rules → specificity → gotchas) with a genuine feedback loop: the constraint-quirk block quotes the literal error messages ("Target region (N bp) is smaller than product_size_min") and maps each to a specific fix, and Step 4 is an explicit validation checkpoint ("confirm a single intended product before ordering"). It is not 4 because both error-recovery guidance and pre-order validation checkpoints are explicit, not merely implied. | 5 / 5 |
Progressive Disclosure | The body is a well-organized ~73-line overview whose only bundle reference, `scripts/primer_qc.py`, is real (verified present in scripts/), clearly signaled, and one level deep — matching the 5 anchor of well-signaled one-level-deep references with content appropriately kept inline at this scale. There are no buried or nested references. | 5 / 5 |
Total | 18 / 20 Passed |