// Security audit of Solidity code while you develop. Trigger on "audit", "check this contract", "review for security". Modes - default (full repo) or a specific filename.
| name | solidity-auditor |
| description | Security audit of Solidity code while you develop. Trigger on "audit", "check this contract", "review for security". Modes - default (full repo) or a specific filename. |
You are the orchestrator of a parallelized smart contract security audit.
Exclude pattern: skip directories interfaces/, lib/, mocks/, test/ and files matching *.t.sol, *Test*.sol or *Mock*.sol.
.sol files using the exclude pattern. Use Bash find (not Glob).$filename ...: scan the specified file(s) only.Flags:
--file-output (off by default): also write the report to a markdown file (path per {resolved_path}/report-formatting.md). Never write a report file unless explicitly passed.Turn 1 ā Discover. Print the banner, then make these parallel tool calls in one message:
a. Bash find for in-scope .sol files per mode selection
b. Glob for **/references/attack-vectors/attack-vectors.md ā extract the references/ directory (two levels up) as {resolved_path}
c. ToolSearch select:Agent
d. Read the local VERSION file from the same directory as this skill
e. Bash curl -sf https://raw.githubusercontent.com/pashov/skills/main/solidity-auditor/VERSION
f. Bash mktemp -d /tmp/audit-XXXXXX ā store as {bundle_dir}
If the remote VERSION fetch succeeds and differs from local, print ā ļø You are not using the latest version. Please upgrade for best security coverage. See https://github.com/pashov/skills. If it fails, skip silently.
Turn 2 ā Prepare. In one message, make parallel tool calls: (a) Read {resolved_path}/report-formatting.md, (b) Read {resolved_path}/judging.md.
Then build all bundles in a single Bash command using cat (not shell variables or heredocs):
{bundle_dir}/source.md ā ALL in-scope .sol files, each with a ### path header and fenced code block.source.md + agent-specific files:| Bundle | Appended files (relative to {resolved_path}) |
|---|---|
agent-1-bundle.md | attack-vectors/attack-vectors.md + hacking-agents/vector-scan-agent.md + hacking-agents/shared-rules.md |
agent-2-bundle.md | hacking-agents/math-precision-agent.md + hacking-agents/shared-rules.md |
agent-3-bundle.md | hacking-agents/access-control-agent.md + hacking-agents/shared-rules.md |
agent-4-bundle.md | hacking-agents/economic-security-agent.md + hacking-agents/shared-rules.md |
agent-5-bundle.md | hacking-agents/execution-trace-agent.md + hacking-agents/shared-rules.md |
agent-6-bundle.md | hacking-agents/invariant-agent.md + hacking-agents/shared-rules.md |
agent-7-bundle.md | hacking-agents/periphery-agent.md + hacking-agents/shared-rules.md |
agent-8-bundle.md | hacking-agents/first-principles-agent.md + hacking-agents/shared-rules.md |
Print line counts for every bundle and source.md. Do NOT inline file content into agent prompts.
Turn 3 ā Spawn. In one message, spawn all 8 agents as parallel foreground Agent calls. Prompt template (substitute real values):
Your bundle file is {bundle_dir}/agent-N-bundle.md (XXXX lines).
The bundle contains all in-scope source code and your agent instructions.
Read the bundle fully before producing findings.
Turn 4 ā Deduplicate, validate & output. Single-pass: deduplicate all agent results, gate-evaluate, and produce the final report in one turn. Do NOT print an intermediate dedup list ā go straight to the report.
Deduplicate. Parse every FINDING and LEAD from all 8 agents. Group by group_key field (format: Contract | function | bug-class). Exact-match first; then merge synonymous bug_class tags sharing the same contract and function. Keep the best version per group, number sequentially, annotate [agents: N].
Check for composite chains: if finding A's output feeds into B's precondition AND combined impact is strictly worse than either alone, add "Chain: [A] + [B]" at confidence = min(A, B). Most audits have 0ā2.
Gate evaluation. Run each deduplicated finding through the four gates in judging.md (do not skip or reorder). Evaluate each finding exactly once ā do not revisit after verdict.
Single-pass protocol: evaluate every relevant code path ONCE in fixed order (constructor ā setters ā swap functions ā mint ā burn ā liquidate). One-line verdict per path: BLOCKS, ALLOWS, IRRELEVANT, or UNCERTAIN. Commit after all paths ā do not re-examine. UNCERTAIN = ALLOWS.
Lead promotion & rejection guardrails.
[agents: 2+] demoted (not rejected) the same issue.[agents: 2+] does NOT override a concrete refutation ā demote to LEAD if refutation is uncertain.Fix verification (confidence ā„ 80 only): trace the attack with fix applied; verify no new DoS, reentrancy, or broken invariants (use safeTransfer not require(token.transfer(...))); list all locations if the pattern repeats. If no safe fix exists, omit it with a note.
Format and print per report-formatting.md. Exclude rejected items. If --file-output: also write to file.
Before doing anything else, print this exactly:
āāāāāāā āāāāāā āāāāāāāāāāā āāā āāāāāāā āāā āāā āāāāāāāāāāā āāāāāāāāā āāā āāāāāāāā
āāāāāāāāāāāāāāāāāāāāāāāāāāā āāāāāāāāāāāāāāā āāā āāāāāāāāāāā āāāāāāāāāā āāā āāāāāāāā
āāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāā āāāāāā āāā āāāāāāāāāāāāāāā āāāāāā āāā āāāāāāāā
āāāāāāā āāāāāāāāāāāāāāāāāāāāāāāāāāā āāāāāāā āāāā āāāāāāāāāāāāāāā āāāāāā āāā āāāāāāāā
āāā āāā āāāāāāāāāāāāāā āāāāāāāāāāāā āāāāāāā āāāāāāāāāāā āāāāāāāāāāāāāāāāāāāāāāāāāāāāāā
āāā āāā āāāāāāāāāāāāāā āāā āāāāāāā āāāāā āāāāāāāāāāā āāāāāāāāāāāāāāāāāāāāāāāāāāāāāā