Getting it into your agent
It runs from inside its repository, so the clone comes first — what it calls does not travel with the file alone.
git clone --depth 1 https://github.com/ShulkwiSEC/bb-hugenpx agentmods add skills/shulkwisec/bb-huge/csptWrote 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/shulkwisec/bb-huge/cspt)<a href="https://agentmods.dev/skills/shulkwisec/bb-huge/cspt"><img src="https://agentmods.dev/badge/skills/shulkwisec/bb-huge/cspt/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/shulkwisec/bb-huge/cspt"><img src="https://agentmods.dev/badge/skills/shulkwisec/bb-huge/cspt.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 | $0.00114 | $0.02129 |
| Opus 5 | $0.00057 | $0.01064 |
| Sonnet 5 | $0.00023 | $0.00426 |
| Haiku 4.5 | $0.00011 | $0.00213 |
Grade A, and why
cspt scanned grade A with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 7d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
grep -rn 'fetch(\|xhr.open\|axios.get\|axios.post' src/ | grep '+' How it starts
The opening of the file, as written. The whole thing — 163 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Client-Side Path Traversal (CSPT)
What Is Broken and Why
Client-Side Path Traversal occurs when attacker-controlled input is concatenated
directly into the path component of a JavaScript fetch() or XHR URL without
proper encoding. The injected ../ sequences traverse the URL path, redirecting
the request to an unintended endpoint. Unlike server-side path traversal (which
reads files), CSPT redirects API calls — enabling response injection, data exfiltration,
and XSS when chained with an open redirect that fetch() auto-follows to an
attacker-controlled domain.
Key Signals
- JavaScript source with user-controlled input concatenated into a fetch/XHR path:
fetch("/api/users/" + userId + "/profile") fetch(`/api/posts/${postId}/comments`) xhr.open("GET", "/api/data/" + param) - Input lands in the path segment (before
?), not the query string - No
encodeURIComponent()wrapping the input, or encoding applied only to specific chars - Application has open redirect endpoints (
/redirect?u=,/oauth/authorize?redirect_uri=) - SPA frameworks where routing inputs feed directly into API calls
- Response content is rendered as HTML or passed to
innerHTML/eval()
Methodology
- Map all
fetch()/XHRcalls in JavaScript source — look for string concatenation in the URL path. - Identify which parameters control the path segment (not query string).
- Test basic traversal: inject
../../../anything— did the request path change? - Determine depth needed: count path segments to traverse to root or to a useful endpoint.
- Find a usable gadget endpoint — open redirect (
/redirect?u=ATTACKER) or any endpoint whose response is reflected in the DOM. - Craft final payload: traverse to the gadget, chain to attacker-controlled response.
- If WAF blocks, apply encoding bypass strategy (see table below).
- Confirm XSS or data exfiltration via OOB callback.
WAF Bypass: Encoding Level Matrix
| Situation | Strategy | Example |
|---|---|---|
| No WAF | Use plain ../ |
../../../gadget |
| WAF level = App level | Encode dots: %2e%2e/ |
%2e%2e/%2e%2e/%2e%2easdf |
| WAF level < App level | Over-encode slashes | ..%252f..%252f..%252fasdf |
| WAF level > App level | Pad with dummy segments | a%252fa%252fa%252fa/../../../../asdf |
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.
- 7d ago First seen · 163 lines · 114 tokens per session scan A 97aeff354b03
cspt is a skill published in the GitHub repository ShulkwiSEC/bb-huge (22 stars, last pushed 2mo ago), licensed MIT. It adds 114 tokens to every session and 2,129 once invoked, about $0.0006 per session on Opus 5. A static security scan graded it A with 1 finding (makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-03.
Other skills, from other repositories
continuum-streaming
Stream tokens, tool calls, handoffs, and memory events out of a Continuum agent in real time using runner.runstream() and the EventType enum. Invoke when the user asks "stream tokens to UI", "websocket chat", "live progress", "see tool execution as it happens", or anything that needs token-by-token output.
scaffold-vite-react
Scaffold Vite + React POC frontend in delivery worktree output directory.
skill-extract
Reverse-engineer design systems, tokens, and components from live products or screenshots.
agent-frontend-standard
A set of engineering guidelines for building consistent front-end interfaces with AI agents. It covers shared design components, type-checked interfaces, visual checks and staged testing requirements.
design-system
Design system management for frontend agents. Actions: init (create DESIGN.md from template), load (inject into agent context), validate (check component compliance). Use when: (1) starting a frontend project, (2) generating UI components, (3) reviewing frontend code for design consistency. Triggers: /design-system…
perf
Performance optimization skill. Core Web Vitals via Lighthouse, bundle size analysis, metrics tracking over time. Use when: (1) optimizing frontend performance, (2) analyzing bundle size, (3) tracking metrics regression. Triggers: /perf, 'performance audit', 'core web vitals', 'bundle size'.