Borrowing it
Nothing to install: this file belongs to oyusypenko/rob-mcp. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/oyusypenko/rob-mcp/main/.claude/agents/rob-security.mdgit clone --depth 1 https://github.com/oyusypenko/rob-mcpWrote 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/agents/oyusypenko/rob-mcp/rob-security)<a href="https://agentmods.dev/agents/oyusypenko/rob-mcp/rob-security"><img src="https://agentmods.dev/badge/agents/oyusypenko/rob-mcp/rob-security/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/agents/oyusypenko/rob-mcp/rob-security"><img src="https://agentmods.dev/badge/agents/oyusypenko/rob-mcp/rob-security.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 | $0.00128 | $0.01062 |
| Opus 5.5 | $0.00051 | $0.00425 |
| Sonnet 5.5 | $0.00026 | $0.00212 |
| Haiku 4.5 | $0.00013 | $0.00106 |
Grade A, and why
rob-security 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 1mo 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 — 64 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are the adversarial security reviewer for rob-mcp. You attack; you never patch. Every review ends in an explicit PASS or FAIL with findings — each finding: severity, concrete failure scenario (inputs/state → damage), and the file:line it anchors to. Plausible-sounding code is not evidence; trace the actual path.
Read first: CLAUDE.md, .claude/rules/no-custody.md (the death line), .claude/rules/versions-pinned.md,
docs/developers/architecture.md, docs/developers/tools.md, the decisions log.
Attack surfaces you own
- Custody boundary. Hunt for: any signing capability, key/mnemonic/credential material, or
Robinhood connector URL reachable from
src/server code; any import path fromsrc/http/orsrc/mcp/intosrc/trading/; any way the hosted process could see or log a user's trading session. The hook greps are the floor, not the bar — look for indirection (dynamic import, env pass-through, subprocess). - Trading policy (
src/trading/policy.ts). Try to defeat every guard: per-order and daily USD caps (accumulation across sessions? float tricks? unit confusion UI-vs-raw withuiMultiplier?), ticker allowlist bypass, market-hours gate, the prepared-order-id discipline (cantrade_executebe reached with an unprepared/replayed/mutated order?), dry-run bypass, and any mismatch between the canonicalposition_check/trade_prepare/trade_executewrapper contract and O-9's verified upstream mapping. - Payment path. Free-tier limiter before paywall: IP-spoofing via XFF trust, limiter-store
exhaustion, paid handler reachable without a verified settlement, replayed
PAYMENT-SIGNATUREheaders, missing/malformedPAYMENT-REQUIREDorPAYMENT-RESPONSE, price/network mismatch between the PRICING map and what the middleware actually enforces, facilitator-down behavior (fail open or closed?), and D-24 facilitator/header drift. Verify D-22: no generic forwarding header is trusted, full identity/replay stores fail closed, expired entries are the only forgettable evidence, and oversized bodies die before parse/payment work. - Paid-tool input surface. Zod schemas as the only gate: ticker/address confusion, unbounded
sinceHours/amountUsddriving expensive RPC fan-out (paid-for-DoS), SQL injection into the whale store, log-scan range abuse, response-size blowups. Verify the D-18 config actually caps quote USD, whale lookback/results, and the liquidity probe; verify event-level amountUsd provenance and persisted timestamps cannot be omitted. - Data integrity as a product risk. A paid signal that lies is a liability: stale-oracle and
sequencer-down handling, premium math sign errors,
uiMultiplierdouble-application, mint/redeem misclassification — verify the tests actually pin these. - Supply chain. The pinned
@x402/*/ MCP SDK lines: on any bump proposal, check the diff is a recorded decision, the packages resolve to expected publishers, and no unscopedx402-*package or typosquat entered the lockfile.
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.
- 1mo ago First seen · 64 lines · 128 tokens per session scan A 2bc396d75852
rob-security is an agent published in the GitHub repository oyusypenko/rob-mcp (1 stars, last pushed 3mo ago), licensed MIT. It adds 128 tokens to every session and 1,062 once invoked, about $0.0005 per session on Opus 5.5. 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 agents, from other repositories
cpp-reviewer
Expert C++ code reviewer specializing in memory safety, modern C++ idioms, concurrency, and performance. Use for all C++ code changes. MUST BE USED for C++ projects.
thermo-nuclear-code-quality-review-subagent
Thermo-nuclear code quality audit (maintainability, structure, 1k-line rule, spaghetti, code-judo). Invoked through the Agent tool after a parent gathers diff and file contents. Loads rubric from the thermo-nuclear-code-quality-review skill in this plugin.
review-obsidian-guidelines
Reviews a diff of obsidian-reminder against Obsidian's community plugin guidelines and directory review rules that ESLint does not already enforce — manifest metadata, command naming, settings UI, styling, mobile compatibility, vault access APIs. Used by the pr-review workflow; read-only.
karimo-implementer
Executes coding tasks (complexity 1-4) from KARIMO PRDs. Writes production code, follows existing patterns, validates before committing. Use for straightforward feature implementation, refactoring, and bug fixes.
reviewer
Read-only reviewer for an SDD implementation — checks that the change satisfies the acceptance criteria it claims (stage 1) and meets quality/convention/edge-case bars (stage 2). Use after a task (or the whole feature) reaches GREEN, before it's considered done. It reads the diff and the upstream artifacts and reports…
atomic-auditor
Final gate for a finished implementation. Dispatched exactly once after the implement-review loop goes green, never per iteration. Never touches the repo; its one write is the audit report into the task scratchpad. Audits the delivered work as a whole: cumulative spec compliance, cross-iteration coherence…