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.
npx skills add UnboundCompute/security-agent-skills --skill auditing-csrf-and-unsafe-state-changesgit clone --depth 1 https://github.com/UnboundCompute/security-agent-skillsWrote 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/skills/unboundcompute/security-agent-skills/auditing-csrf-and-unsafe-state-changes)<a href="https://agentmods.dev/skills/unboundcompute/security-agent-skills/auditing-csrf-and-unsafe-state-changes"><img src="https://agentmods.dev/badge/skills/unboundcompute/security-agent-skills/auditing-csrf-and-unsafe-state-changes/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/skills/unboundcompute/security-agent-skills/auditing-csrf-and-unsafe-state-changes"><img src="https://agentmods.dev/badge/skills/unboundcompute/security-agent-skills/auditing-csrf-and-unsafe-state-changes.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 | $0.00161 | $0.02209 |
| Opus 5.5 | $0.00064 | $0.00884 |
| Sonnet 5.5 | $0.00032 | $0.00442 |
| Haiku 4.5 | $0.00016 | $0.00221 |
Grade A, and why
auditing-csrf-and-unsafe-state-changes 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 — 142 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Auditing CSRF and unsafe state changes: when the victim's cookies do the attacker's work
A browser attaches a site's cookies to every request to that site, whoever caused the request. So when a state-changing endpoint authorizes an action on the session cookie alone, an attacker's page can submit a form or fire a request to that endpoint, the victim's browser attaches the session, and the change happens as the victim without them ever intending it. The defense is proof that the request originated from the app's own pages and not a foreign one: an unpredictable token bound to the session that an attacker cannot guess or read cross-origin, or a cookie policy that withholds the cookie on cross-site requests. The bug is a state change that trusts the cookie and asks for nothing else. You find it by listing every action that mutates state and checking what, beyond the cookie, each one requires.
When to use
- An endpoint changes data, settings, access, or funds and is authorized by a session cookie.
- Forms or actions rely on a token, and you need to confirm it is present, validated, and session-bound.
- The app relies on same-site cookie behavior, custom headers, or content-type as its only cross-site guard.
Scope check
Test cross-site request forgery only against applications you own or are authorized to assess, using test accounts and a benign state change you can observe and reverse, never triggering a real irreversible action on another user. A confirmed case performs a real mutation as the victim, so stay within the authorized scope. If you can't name the authorization, stop.
The loop
- Establish what each state change requires beyond the cookie first. For every mutating endpoint, determine whether it demands an unpredictable, session-bound token that it validates, or relies on a fully-enforced same-site cookie, or nothing at all. This is the false-positive killer: an endpoint that validates a session-bound token on every mutation, or is reached only with a cookie the browser withholds cross-site, is defended, while one authorized on the ambient cookie alone is not. Name the requirement before building a cross-site request.
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 · 142 lines · 161 tokens per session scan A 8c2a26a966f4
auditing-csrf-and-unsafe-state-changes is a skill published in the GitHub repository UnboundCompute/security-agent-skills (5 stars, last pushed 1mo ago), licensed MIT. It adds 161 tokens to every session and 2,209 once invoked, about $0.0006 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 2026-09-05.
Other skills, from other repositories
prodcheck-review
Review this codebase against the prodcheck pre-production checklists — security, performance, scale, integrations, post-launch readiness and people. Use when asked to check whether a project is ready to ship, to audit an area before launch, or to work through a specific checklist. Produces evidence with file:line…
integrate-arcjet-guard-genkit
Integrate Arcjet security into a Genkit JS agent using @arcjet/guard — wrap ai.defineTool, put guardMiddleware on generate({ use }) for unwrapped / MCP / filesystem tools, and read a caller-owned id from generate({ context }). Use when asked to add Arcjet to genkit, rate limit its tools, screen inbound messages, or…
integrate-arcjet-guard-tanstack-ai
Integrate Arcjet security into a TanStack AI chat() app using @arcjet/guard — put guardMiddleware first on chat({ middleware }) so onBeforeToolCall gates tools, and read a caller-owned id from chat({ context }). Use when asked to add Arcjet to TanStack AI, rate limit its tools, screen inbound messages, or block prompt…
integrate-arcjet-guard-claude-managed-agents
Integrate Arcjet security into Claude Managed Agents (hosted REST+SSE, beta managed-agents-2026-04-01) using @arcjet/guard — screen user.message / initialevents before sessions.events.send, and gate custom tools on agent.customtooluse. Use when asked to add Arcjet to Claude Managed Agents, rate limit custom tools, or…
linear-claude-skill
Manage Linear issues, projects, and teams.
varlock-claude-skill
Secure environment variable management ensuring secrets are never exposed in Claude sessions, terminals, logs, or git commits.