Content
68%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-organized, mostly executable offensive-security reference that assumes Claude's competence and avoids padding, but it lacks explicit validation/verification checkpoints for a destructive workflow and a few attack classes are described only in prose.
Suggestions
Add explicit validation checkpoints to the workflow (e.g. 'Verify the fork reproduces current governance state before attacking', 'Confirm quorum/votingPeriod via recon before simulating').
Flesh out attack classes 3, 4, and 6 with concrete executable snippets instead of bullet-only descriptions.
Move external post-mortem/audit links into a references/ bundle file or mark them clearly as external so the SKILL.md stays a lean overview.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is lean and assumes Claude knows Solidity/DeFi (no padding explaining what a DAO or flash loan is); the Beanstalk anecdote earns its place, though a few explanatory prose lines could be trimmed. | 4 / 5 |
Actionability | Executable `cast call`/`cast logs`/`forge inspect` recon commands and Solidity snippets are concrete, but attack classes 3, 4, and 6 are bullet-only without complete executable code, leaving minor gaps. | 4 / 5 |
Workflow Clarity | A rough sequence exists (identify attack class → recon → simulate on fork) but there are no validation/verification checkpoints; because this is a destructive offensive skill, workflow clarity is capped at 3 per the rubric. | 3 / 5 |
Progressive Disclosure | Content is organized into clear single-file sections (Attack classes, Recon, Defender's checklist, Tooling, References) with no nested or buried references; no bundle files exist, so structure is judged on section organization, which is good with minor gaps. | 4 / 5 |
Total | 15 / 20 Passed |