Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
git clone --depth 1 https://github.com/omermaksutii/RugProofWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/commands/omermaksutii/rugproof/replay-incident)<a href="https://agentmods.dev/commands/omermaksutii/rugproof/replay-incident"><img src="https://agentmods.dev/badge/commands/omermaksutii/rugproof/replay-incident/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/commands/omermaksutii/rugproof/replay-incident"><img src="https://agentmods.dev/badge/commands/omermaksutii/rugproof/replay-incident.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00024 | $0.00729 |
| Opus 5 | $0.00012 | $0.00365 |
| Sonnet 5 | $0.00005 | $0.00146 |
| Haiku 4.5 | $0.00002 | $0.00073 |
Grade A, and why
replay-incident scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 10d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 91 lines — stays where its author put it; the contents beside it link to each section on GitHub.
/replay-incident — replay a historical hack
Pulls a real exploit tx from chain, replays it on a fork, and walks through what happened.
Great teaching tool. Also great content — every replay can become a tweet or blog post.
Procedure
Step 1 — Fetch the tx
mcp__block-explorer__get_tx(chain=<chain>, hash=<hash>)
mcp__block-explorer__get_trace(chain=<chain>, hash=<hash>)
Pull the calldata, value, from/to, internal calls, state diff, gas, and block.
Step 2 — Fork just before the tx
mcp__anvil__fork(chain=<chain>, block=<tx_block - 1>)
This gives us the pre-attack state.
Step 3 — Identify the protocol and contracts touched
Pull the verified source of every contract in the trace.
Step 4 — Replay
mcp__anvil__send_raw_tx(<original-tx>)
Capture the actual state changes; compare to the live chain post-tx data — should match.
Step 5 — Annotate the trace
Walk through each internal call:
Block 18,234,567 — Curve Finance exploit (2023-07-30)
External call: 0xattacker → 0xfraxusdc_pool.remove_liquidity_one_coin(...)
│
├─ Internal: pool transfers $1M USDC to attacker
│ Then calls pool.balanceOf(attacker) for reward accounting...
│ │
│ └─ READ-ONLY REENTRANCY: pool state is mid-update during this read
│
└─ attacker contract uses the stale read to mint reward shares
based on the larger-than-real balance
Step 6 — Vuln-class mapping
Map the incident to the [[skills]] taxonomy:
Root cause: read-only reentrancy (skills/reentrancy)
Amplifier: reward accounting based on token.balanceOf during mid-update
Lesson: for any oracle/getter call from a reentrant contract, add
ReentrancyGuard or rely on a settled-state snapshot.
Step 7 — Output
The full walkthrough + a "what to learn" section + (if --card flag) a PNG card for sharing.
Notable replayable incidents to demo
- 2022-04-17: Beanstalk governance flash-loan ($182M)
- 2022-08-01: Nomad bridge replay attack ($190M)
- 2023-07-30: Curve / Vyper compiler reentrancy ($73M)
- 2023-11-22: KyberSwap concentrated liquidity precision ($55M)
- 2024-03-13: Munchables blacklist-able EOA admin ($63M)
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 10d ago First seen · 91 lines · 24 tokens per session scan A 661b809422dd
replay-incident is a command published in the GitHub repository omermaksutii/RugProof (9 stars, last pushed 1mo ago), licensed MIT. It adds 24 tokens to every session and 729 once invoked, about $0.0001 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other commands, from other repositories
krait-poc
Write and run a valid Foundry proof-of-concept that proves (or disproves) a Solidity exploit by asserting the actual harm on a forked chain or against local source.
verify-e2e-flow
Verify that the E2E scripts for the specified chain follow the documented transaction flow.
watch-wallet
Watch Wallet operates online and manages public keys only. It creates unsigned transactions, sends signed transactions, and monitors transaction status.
keygen-wallet
Keygen Wallet operates offline as a cold wallet. It generates keys, creates multisig addresses, and provides the first signature for multisig transactions.
fix-btc-e2e-p11
Fix errors in BTC E2E test (Pattern 11: P2TR Tapscript M-of-N).
fix-btc-e2e-p2
Fix errors in BTC E2E test (Pattern 2: P2PKH 2-of-3 Multisig).