Content
78%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.
The body delivers genuinely specialized operational knowledge in highly concrete, actionable form — thresholds, deadlines, escalation triggers, and templates throughout. Its weaknesses are structural: everything is inlined in one long file with no reference bundle, and the workflow's validation steps are implicit rather than explicit.
Suggestions
Move the dense reference material — Carrier Behaviour by Mode, Seasonal and Cyclical Patterns, Fraud and Red Flags — into one-level-deep reference files (e.g., references/carrier-behaviors.md) linked from short summary bullets in SKILL.md, keeping the main file as an overview.
Trim the role-play preamble and per-section intro sentences (e.g., 'Understanding how different carrier types operate changes your resolution strategy') to reduce token load without losing operational substance.
Add explicit validation checkpoints to the How It Works sequence — e.g., verify documentation is complete against the carrier-specific requirements list before filing, and confirm filing date is inside the 9-month Carmack window before submission.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Mostly efficient: nearly every line carries domain-specific data Claude would not reliably know (dollar thresholds, the 9-month Carmack window, 30/120-day carrier response deadlines, settlement floors). Not 5 because of the role-play preamble ('You are a senior freight exceptions analyst with 15+ years...'), per-section intro sentences ('Understanding how different carrier types operate changes your resolution strategy'), and an Examples section that overlaps with Key Edge Cases; not 3 because padding is minor relative to the dense operational content. | 4 / 5 |
Actionability | Fully actionable instruction-only guidance: exact dollar tiers with prescribed actions ('Reject settlements below 80%'), precise deadlines (5-day concealed damage, 14/21-day Montreal Convention notice), an escalation table with hour-level timelines, copy-paste-able subject-line templates ('ESCALATION: Unresolved Exception — {shipment_ref} — {days} Days'), and quoted example phrasings. Specific examples cover the common cases (damage claim, detention dispute, lost shipment). | 5 / 5 |
Workflow Clarity | The 5-step How It Works sequence (classify → apply workflow → document → escalate → file/track) is clear, and checkpoints exist via the three-axis severity assessment, the automatic escalation trigger table, and time-based re-checks ('Escalate at 30-day mark if no resolution'). Not 5 because validation is implicit — there is no explicit verify step such as confirming documentation completeness against carrier requirements before filing a claim; not 3 because sequencing and most checkpoints are present and unambiguous. | 4 / 5 |
Progressive Disclosure | No bundle files exist (no references/, scripts/, or assets/), and all ~220 lines live inline in SKILL.md with good headers but content that should be separate is inline — Carrier Behaviour by Mode, Seasonal Patterns, Fraud Red Flags, and the edge-case playbooks are exactly the material that belongs in one-level-deep reference files. Not 2 because the section structure is clear and navigable, not monolithic; not 4 because the body far exceeds overview size and nothing is split out, even though the skill itself hints at it ('Brief summaries are included here so you can expand them into project-specific playbooks'). | 3 / 5 |
Total | 16 / 20 Passed |