Content
71%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 is an efficient, highly actionable checklist-style audit guide with concrete greps, kill criteria, and a working scanner integration, held back by a semi-pseudocode PoC template and a reference section whose target files are not present in the bundle. Structure and sequencing are good but not fully validated end-to-end.
Suggestions
Ship the referenced files inside the skill bundle (e.g. references/meme-coin-bugs.md, references/solana-token-audit.md, references/dex-lp-attacks.md, and a real scripts/token_scanner.py) and update the body's paths so every reference resolves.
Complete the Foundry PoC template so it is runnable — replace the '// ... buy tokens on Uniswap' and '// Assert token price crashed' ellipses with actual swap calls and price-diff assertions.
Add an explicit numbered audit sequence (pre-dive kills → greps per bug class → scanner → PoC → report) so the workflow order is unambiguous rather than implied by section layout.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is dense and assumes competence (no explanations of what blockchains or LPs are), with every bug class compressed to a grep, a stat, and a kill criterion. Minor trimmable padding remains, e.g. the 'THE ONE RULE' pull-quote block and narrative comments in the PoC template ('Step 3: Victim's tokens are now worthless'). | 4 / 5 |
Actionability | Grep patterns, scanner invocations ('python3 bughunter/tools/token_scanner.py contracts/Token.sol'), and Solana CLI checks are copy-paste ready. However, the Foundry PoC template contains placeholder ellipses ('// ... buy tokens on Uniswap', '// ... Add liquidity', '// Assert token price crashed') making it a template rather than fully executable code. | 4 / 5 |
Workflow Clarity | A clear sequence is implied and partially explicit — 'Check these BEFORE reading a single line of code' establishes the pre-dive gate, each bug class has a 'Kill if:' checkpoint, and the scanner section ends with what it does and does not check. Not a numbered end-to-end procedure, so it sits just below the fully-checkpointed anchor. | 4 / 5 |
Progressive Disclosure | The closing 'FULL REFERENCE FILES' section is well-signaled with one-line descriptions, but the referenced paths ('web3/10-meme-coin-bugs.md', 'web3/11-solana-token-audit.md', 'web3/12-dex-lp-attacks.md', 'bughunter/tools/token_scanner.py') do not exist in this skill's bundle — no references/, scripts/, or assets/ directories are present — so navigation is broken. Additionally, the ~80-line Foundry PoC template is content that arguably belongs in a referenced file. | 3 / 5 |
Total | 15 / 20 Passed |