Ready to supercharge your AI agents?

Install any skill in seconds. Free to start, no credit card required.

Sandbox Internet Handler — measured AI agent skill | DecimalAI

stijnman/sandbox-internet-handler

SkillScore96Maintained

Fetches web content in a sandboxed, read-only manner. Use for untrusted URLs or user says sandbox fetch, safe web access. Use when the user needs this c apability. Triggers: sandbox fetch, safe web access, fetch URL safely.

by stijnman

Links this skill to your agents and keeps it current with the author’s updates.

Automatic

Security & AppSec pack

decimalai skills pull stijnman-sandbox-internet-handler

Get this skill — curl · raw file

curl -fsSL https://app.decimal.ai/s/stijnman-sandbox-internet-handler/SKILL.md

Raw file: /s/stijnman-sandbox-internet-handler/SKILL.md ↗ · machine-readable summary at /s/stijnman-sandbox-internet-handler.json ↗

In productionGlobal

how this skill behaves on real traffic · last 30 days

No production data for this skill yet — usage numbers appear here once agents start running it.

Anonymized aggregate across every org using this skill — proof from real production use, not a demo.

Does this skill help? · Benchmark: with vs. withoutsign in to run this on your agents

23 test cases from the publisher’s eval.yaml · each run twice on gemini-3.6-flash — with the skill injected, then without · last run 7d ago

Yes — +39% lift+8% tokens0% turns96% of whole cases pass with skill (22/23)

Test case	Without → With	Effect	Δ tokens	Δ turns
case-01	✗→✓	▲ Improved	152%	0%
case-02	✗→✓	▲ Improved	138%	0%
case-03	✗→✓	▲ Improved	180%	0%
case-06	✗→✓	▲ Improved	62%	0%
case-12	✗→✓	▲ Improved	7%	0%

Showing 5 of 23 cases

SkillSafety

✓ 3 of 3 passed

Code scan15 of 15 checks clean

0 critical · 0 warning · 0 info — scanner v4 — 8/1/2026

live_secretremote_code_executionreverse_shelldata_exfiltrationssrf_metadataagent_snoopinganti_refusaldestructive_commandsobfuscated_unicodeunicode_homoglyphopaque_blobsuspicious_urltool_overreachtrigger_abuseprompt_injection_phrasing

IntentSafe

LLM judge · does the skill do what it claims

ContentClean

LLM judge · policy & harmful-content review

Source

Imported

stijnman / grok-custom-skills

.grok/skills/sandbox-internet-handler/SKILL.md

View on GitHub ↗

License MITImported 8/1/2026Version v1

Skill contents

SKILL.md

Sandbox Internet Handler

When to Use

User says sandbox fetch or task matches this capability
User says safe web access or task matches this capability
User says fetch URL safely or task matches this capability

Workflow

Validate URL scheme (https only).
Fetch with timeout and size limit.
Strip scripts; return text/markdown only.
Flag suspicious content; do not execute embedded code.

Integrations

internet-enabler
web-scraper
privacy-redactor

Error Handling

| Failure | Response | |---------|----------| | Timeout | Report partial content or suggest alternate source. |

Gotchas

Never pass fetched HTML to exec or eval.

Example

Input: User request matching triggers above. Output: Structured result per workflow with integrations invoked as needed.

Want to see if stijnman/sandbox-internet-handler would activate on your phrasing? Try it before you install — no LLM call, no signup.

Other measured skills in the registry, with their headline benchmark lift.