How do I install the Hunt File Upload skill?

Run npx skillmds@latest add duriantaco/hunt-file-upload in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

Is Hunt File Upload safe to use?

This skill has not completed SkillMD's automated safety review yet. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Hunt File Upload?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is Hunt File Upload free to use?

Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.

Who published Hunt File Upload?

duriantaco (@duriantaco) published this skill. Their other Agent Skills are listed on their SkillMD profile.

Hunt File Upload by duriantaco — Agent Skill

Name: Hunt File Upload
Author: duriantaco

Hunt File Upload

Investigate authorized file-upload, stored-file, and upload-parser boundaries. Use when typed discovery identifies upload forms, multipart requests, filename or content-type handling, stored-file readback, image or document ingestion, archive processing, XML or SVG parsing, or suspected upload-policy bypasses.

duriantaco b981182 2 files · 1.9 KB Updated Sep 17, 2026

File contents

Hunt File Upload Boundaries

Treat this card as advisory prioritization. Expected acceptance and returned storage paths are not vulnerabilities by themselves.

Workflow

Preserve the upload method, identity, form state, media policy, and returned storage reference.
file_fetch_parser includes fixed interpreted-file and deserialization candidates. Run it only when the complete set is authorized; preserve multipart shape and change one property.
Follow only target-returned same-origin readback. Use xxe_boundary only for a typed XML or SVG parser, and keep acceptance, storage, serving, readback, and parser effects separate.
Preserve controls and evidence references. Report validated vulnerabilities independently of any challenge objective.

Evidence Gate

Confirm only a replayable policy bypass with security-relevant target-origin readback/effect, or an independently validated parser flaw. Acceptance, extension or media disagreement, and saved paths remain observations. Keep any contract_missing result suspected rather than promoting it.

Stop Conditions

Stop without a replay, required state, in-scope readback, clear control, or policy budget. Do not add overwrites, persistent shells, decompression bombs, or parser exhaustion; clean up canaries.

Frequently asked questions