A command that checks an iOS app for security weaknesses in stored data, network traffic, encryption, runtime behavior, and authorization. It uses Frida, a tool for inspecting and changing a running app.

Use it to inspect secrets and tokens, test network requests, try attack inputs such as SQL injection and cross-account access, examine deep links, and check whether paid features are enforced by the server.

It organizes many vulnerability checks and shows when human interaction is needed because the testing tool cannot operate the app's user interface.

hunt — Command for Claude Code by xtofuub: A command that checks an iOS app for security…

Install

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.

agentmods

npx agentmods add commands/xtofuub/frida-mcp-server/hunt

Clone the repo

git clone --depth 1 https://github.com/xtofuub/frida-mcp-server

Made for: Claude Code.

Wrote 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.

README.md

[![agentmods](https://agentmods.dev/badge/commands/xtofuub/frida-mcp-server/hunt.svg)](https://agentmods.dev/commands/xtofuub/frida-mcp-server/hunt)

Your own site

<a href="https://agentmods.dev/commands/xtofuub/frida-mcp-server/hunt"><img src="https://agentmods.dev/badge/commands/xtofuub/frida-mcp-server/hunt.svg" alt="Measured on agentmods" height="20"></a>

Per session 22 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.

When invoked 919 The whole file, excluding the scripts and references it only reads on demand.

Security scan A 0 findings. Scan, not verified.

Origin original No closer match found in the catalogue.

Token cost

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	$0.00022	$0.00919
Opus 5	$0.00011	$0.00460
Sonnet 5	$0.00004	$0.00184
Haiku 4.5	$0.00002	$0.00092

Measured 5d ago against content hash a3220ebdc5f5, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-05, from the pricing page.

Security

Grade A, and why

hunt 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 5d 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.

.claude/commands/hunt.md · 59 lines

What it actually says

Hunt $ARGUMENTS. Respect the mode tiers from /autopilot (confirm ACTIVE + BYPASS actions unless --mode yolo). Delegate deep work to the ios-hunt subagent (and ios-runtime for the logic phase). Cover, in order, weighting --focus if given.

No UI control: the MCP can't tap buttons. When a flow needs UI, hook/trace first, then tell the user exactly what to press, wait for confirmation, then drain logs and analyze.

Passive (always):

mcp__frida__scan over captured traffic — plaintext HTTP, JWT alg=none, leaked API keys, CORS, cookie flags, stack traces.
mcp__frida__jwt on captured tokens. mcp__frida__strings / mcp__frida__dump for hardcoded secrets. mcp__frida__crypto_logs for keys/IVs (arm mcp__frida__crypto — BYPASS tier). mcp__frida__memory scan for secrets.
Storage: keychain, defaults, cookies, sqlite_query, files/read — tokens/PII at rest, secrets in UserDefaults vs Keychain.

Active (tiered):

mcp__frida__fuzz ranked params — payload_set ∈ sqli, xss, idor_numeric, idor_uuid, path_traversal, cmd_inj, nosql, auth_bypass, buffer_overflow.
mcp__frida__replay / replay_as — strip/swap auth for BOLA/IDOR/mass-assign.
mcp__frida__race — TOCTOU on state-changing endpoints.
mcp__frida__intercept / intercept_match — rewrite in-flight requests.
mcp__frida__open_url — exercise each deep-link scheme handler.

IAP / paywall / entitlements (tiered): does the server enforce purchases, or does the app trust the client? Flip local state (defaults_set, plist via files/read, keychain), flip the gate (gates→exec), force StoreKit/receipt validators, edit RevenueCat/Adapty caches, or rewrite the entitlement response (intercept_match); then confirm whether paid data is still served. See skills/reverse-engineering-ios-app-with-frida/references/iap-paywall-testing.md.

Runtime logic (tiered — the interesting bugs): delegate to ios-runtime.

gates(app_only=True) ranks BOOL-returning decision methods (by type encoding, not name) + backing ivars. Names only weight the score — don't hardcode them; low-score methods can still be the real gate.
trace a candidate → user drives the flow → trace_logs to see which fire.
exec a return-flip (retval.replace(ptr(1))) or instances+inspect+call one object; re-drive the flow; observe whether capability is gained.
See skills/reverse-engineering-ios-app-with-frida/references/runtime-logic-hunting.md.

Bypass (tiered, defense-quality test):

mcp__frida__ssl_unpin, mcp__frida__jb_bypass — if trivially bypassed, that's a finding about defense strength.

Map every candidate to a control in skills/reverse-engineering-ios-app-with-frida/references/masvs-checklist.md and owasp-mobile-top10.md. Use bugbounty-playbooks.md for concrete chains.

For each hit, log the winning technique: python scripts/memory.py log patterns. Output: candidate findings with evidence + the exact tool calls to reproduce. Hand them to /validate.

Changes

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.

5d ago First seen · 59 lines · 22 tokens per session scan A a3220ebdc5f5

Subscribe to this mod's changes

hunt is a command published in the GitHub repository xtofuub/frida-mcp-server (2 stars, last pushed 2mo ago), licensed MIT. It adds 22 tokens to every session and 919 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.

Security

hunt

Getting it into your agent

Wrote this? Show the measurements

What it costs to keep this loaded

Grade A, and why

Nothing flagged

What it actually says

What this file has done since we first saw it

Other commands, from other repositories

checklist

clarify

specify

converge

implement

analyze