Lean and efficiently organized with no padding or explanations of concepts Claude already knows; every section earns its place (Overview, two tactics, Expected Outcome, Validation).

Provides concrete, specific guidance — numbered steps with exact fire-reduction pattern (100,000 → 50,000 → 30,000) and explicit tier matchups (sacrifice weakest vs. strongest, strongest vs. middle, middle vs. weakest) that are directly actionable.

Multi-step processes are clearly sequenced, and the Validation section adds explicit checkpoints for risky operations (confirm the enemy took the bait, verify the ambush position, confirm tier assignments before committing).

A short, single-purpose skill with no bundle files; the body is well-organized into clear sections, which is appropriate and needs no external references.

Lists multiple concrete actions — 'staged retreat (reducing cooking fires from 100,000 to 30,000)', 'feign mass desertion', and 'asymmetric resource allocation (sacrifice weak vs. strong, deploy strong vs. middle)' — rather than vague language.

Clearly answers what it does (employs staged retreat and asymmetric resource allocation for 2-1 victories) and when to use it via an explicit 'Use when facing a numerically superior enemy or planning ambush warfare' clause.

Includes natural phrases a user might say ('numerically superior enemy', 'ambush warfare'), but coverage of common variations is thin and missing broader terms like 'outnumbered' or 'tactical deception'.

Occupies a clear niche — Sun Bin's military deception tactics — with distinctive triggers unlikely to overlap with other skills.