Content
88%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.
An excellent operational playbook: executable scripts, a clear step sequence, and real validation gates (oracle check, sanity gate) before expensive batch extraction. The only weaknesses are minor verbosity in a few asides and a single-file layout that, while well-organized, is denser than the ideal overview-plus-references split.
Suggestions
Trim the cost-justification asides ('saves ~70% of the round-trip cost', 'Saves you from re-checking which quote variant...') — the scripts and the step sequence already justify the approach.
Consider moving the WAF substitution matrix and the anti-patterns section into a small references/ file to keep SKILL.md as a leaner overview, which would also raise progressive disclosure structure.
EXTRACT_EXPR is referenced in Steps 4-6 but never given a worked example (e.g. the concrete information_schema subquery from Step 2); one fully assembled example would make the extraction scripts immediately runnable end-to-end.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Mostly lean operational prose where the one conceptual explanation (MySQL boolean-as-integer arithmetic) is essential to the technique. Minor trims possible, e.g. 'One script run beats 15 sequential bash calls (saves ~70%...)' and 'Saves you from re-checking which quote variant the WAF blocked this iteration.' | 4 / 5 |
Actionability | Copy-paste-ready bash probe and extraction scripts with env-var parameterization, concrete payloads ("x\"||1=1#"), a bench-tested substitution table, and a per-position extraction loop — specific examples cover the common cases. | 5 / 5 |
Workflow Clarity | Clear sequenced workflow (oracle confirmation → token fingerprint → substitution → hex literals → length discovery → extraction) with explicit validation checkpoints for the batch request loops: the ≤2-minute oracle gate, the filter probe, and the Step 6 sanity gate that aborts before a multi-hundred-request extraction on an empty result. | 5 / 5 |
Progressive Disclosure | Single-file skill with no bundle files; sections are clearly headed and all content is needed inline for the coherent playbook. It exceeds the 'under 50 lines' simple-skill case and the substitution matrix/anti-patterns could arguably live in reference files, so structure is good but not exemplary. | 4 / 5 |
Total | 18 / 20 Passed |