Publisher: OpenAI · Editorial profile by AI Tools Ratings · Review method
Not runtime tested; source-document review only. No security certification.
What task is this skill designed for?
Model how an attacker might cross from an uploaded document into a worker process that can access application data.
What input and output should you expect?
Input: Repository paths, a deployment sketch, upload and authentication flows, data sensitivity and the boundaries the review should cover.
Output: Expected, not independently tested: a Markdown threat model with evidence-backed architecture, abuse paths, priorities and suggested controls.
What does it need to run?
Read access to the relevant source and enough deployment context to distinguish a real trust boundary from a guessed one. The workflow refers to bundled prompts and output guidance that should remain available after installation.
Authored for Codex in the OpenAI skills repository. This directory has not independently installed or executed the skill.
How should you approach installation?
- Read the pinned source and its output-contract reference.
- Supply a bounded repository path and a non-secret architecture sketch.
- Review assumptions with a system owner before treating the resulting priorities as an action plan.
Installation guide: discovery, dependencies and runtime checks
What files, accounts or data can it access?
Reads source and architecture material that may expose sensitive system design. The resulting threat model can itself be sensitive; keep it within the intended review audience.
What does the skill cost?
Source instructions have no stated purchase price; the review consumes agent and engineering time.
What are its limits and alternatives?
A threat model prioritizes possible failure paths; it does not prove exploitability or replace vulnerability testing. Incorrect assumptions about worker isolation, network exposure or authentication can change the whole priority order.
Use the security-best-practices profile for concrete language-level implementation checks. Choose threat modeling when the central question is where trust changes and what an attacker could gain.
What could you ask in a small trial?
Try this with a small sample:
Threat-model the document upload and worker paths only. Identify assets, entry points and trust crossings, cite repository evidence, and list unanswered deployment questions before ranking abuse paths.
How would you verify the result?
- Confirm every named component and trust crossing with a code or configuration reference.
- Ask an owner to validate exposure and identity assumptions.
- Turn the highest-priority mitigation into a testable engineering task with an owner.
Use a baseline and record acceptance evidence before treating a successful task as a reliable workflow.
Which source and license were checked?
Read the reviewed SKILL.md at its pinned commit · Complete reviewed folder
Repository and current publisher guidance
Checked:
Apache-2.0 — identified in skills/.curated/security-threat-model/LICENSE.txt at the reviewed commit. Third-party assets and dependencies may have separate terms. License evidence
Version details and review scope
Source: openai/skills/skills/.curated/security-threat-model
Reviewed commit: 49f948faa9258a0c61caceaf225e179651397431
SKILL.md SHA-256: 1283c0dd62a8104d9edda4583569b5d8510b4ddaa45120687c999250fd96bad2
Source-document review. Source text and stated dependencies were inspected for this profile. Referenced code was not comprehensively audited, no runtime test was performed, and source files are not mirrored here.