diff --git a/.github/SECURITY.md b/.github/SECURITY.md
index fa935b0..16e484b 100644
--- a/.github/SECURITY.md
+++ b/.github/SECURITY.md
@@ -1,13 +1,50 @@
-# Security Policy
+# Security policy
 
-## Supported Versions
+sphinx-argparse-cli is a Sphinx extension that renders the command-line interface an `argparse` parser defines. This
+policy says how to report a security problem and how we decide whether a report is one.
 
-| Version | Supported          |
-| ------- | ------------------ |
-| 1.10 +  | :white_check_mark: |
-| < 1.10  | :x:                |
+## Reporting a vulnerability
 
-## Reporting a Vulnerability
+**Do not open a public issue for a security problem.** Use
+[GitHub's private vulnerability reporting](https://github.com/tox-dev/sphinx-argparse-cli/security/advisories/new). If
+you do not have a GitHub account, or you are a Tidelift subscriber, report through the
+[Tidelift security contact](https://tidelift.com/security) instead; Tidelift coordinates the fix and the disclosure with
+us. If neither works for you, email [gaborjbernat@gmail.com](mailto:gaborjbernat@gmail.com).
 
-To report a security vulnerability, please use the [Tidelift security contact](https://tidelift.com/security). Tidelift
-will coordinate the fix and disclosure.
+Include what the attacker controls and what they gain, the steps or a proof of concept, and the versions and platform
+you tested. Write the report yourself; unreviewed tool or model output takes longer to triage.
+
+## What to expect
+
+Volunteers maintain sphinx-argparse-cli, with no on-call rotation. We aim to acknowledge a report within **5 working
+days** and assess it within **15 working days**. Fixes are developed in a private fork attached to the draft advisory,
+ship in a new release, and the advisory, with a CVE requested through GitHub, goes out with that release. The default
+embargo is **90 days** from the report; we agree any change with the reporter. We credit reporters in the advisory under
+the name they choose, or not at all. There is no bug bounty.
+
+## Supported versions
+
+Only the latest release receives fixes. Confirm the problem reproduces on it before reporting.
+
+## Vulnerability or bug
+
+A report is a vulnerability when an attacker who controls only an input that the [threat model](THREAT_MODEL.md) lists
+as untrusted breaks a guarantee sphinx-argparse-cli makes. A report that first needs control of something the threat
+model trusts, such as the documented project's code or the Sphinx configuration, is a bug: we fix it in the open and
+publish no advisory. Classify by who supplies the dangerous input, not by how bad the outcome sounds.
+
+**In scope**
+
+- The directive importing a module, or calling a function, other than the ones its `:module:` and `:func:` options name.
+- The extension reading or writing files beyond the Sphinx build's own outputs.
+
+**Out of scope**
+
+- Code that runs when the directive imports the module and calls the parser factory: that is the documented project's
+  code, trusted like every module autodoc imports.
+- Markup in help strings: the extension parses them as reStructuredText on purpose, so they are trusted like the code
+  that defines them.
+- Markup the documentation author writes around the directive.
+- Sphinx and docutils behavior.
+
+If you cannot tell which side a finding falls on, report it privately and we will decide.
diff --git a/.github/THREAT_MODEL.md b/.github/THREAT_MODEL.md
new file mode 100644
index 0000000..34b851b
--- /dev/null
+++ b/.github/THREAT_MODEL.md
@@ -0,0 +1,90 @@
+# Threat model
+
+This model says what sphinx-argparse-cli trusts, where untrusted input enters, and which guarantees hold at each
+boundary, so a report can be classified as a vulnerability or a bug. [SECURITY.md](SECURITY.md) covers reporting; if the
+two disagree, SECURITY.md wins and we fix this file.
+
+## How sphinx-argparse-cli works
+
+The `sphinx_argparse_cli` directive imports the module named by `:module:`, gets the parser either by calling the
+function named by `:func:` or, with `:hook:`, by running that function while `ArgumentParser.parse_known_args` and
+`parse_known_intermixed_args` are patched to capture the parser. It walks the parser's actions and subcommands and emits
+docutils nodes for them, parsing help text as reStructuredText, and writes its stylesheet into the build output.
+
+```mermaid
+flowchart LR
+    conf[("conf.py and .rst<br/>(trusted)")] --> ext["sphinx-argparse-cli"]
+    ext -->|"import :module:, call :func:"| code["Project code<br/>(trusted)"]
+    code -->|"ArgumentParser: help, choices"| ext
+    ext -->|"docutils nodes"| html["Built docs"]
+    classDef trusted fill:#cfe4ff,stroke:#1f6feb,color:#0b1f3a;
+    classDef boundary fill:#ffe3a3,stroke:#d29200,color:#3a2c00;
+    class conf,ext,code trusted;
+    class html boundary;
+```
+
+## Compared with similar tools
+
+We drew the boundary where comparable projects in other ecosystems draw theirs. The shared rule, stated by the
+[Node.js threat model](https://github.com/nodejs/node/blob/main/SECURITY.md#the-nodejs-threat-model), is that a report
+which first needs a trusted element compromised is out of scope; the first triage question, as
+[DOMPurify](https://github.com/cure53/DOMPurify/blob/main/SECURITY.md#what-we-consider-a-vulnerability) puts it, is what
+the attacker controls.
+
+| Project                                                                              | Language   | Where it draws the line                                                                                             |
+| ------------------------------------------------------------------------------------ | ---------- | ------------------------------------------------------------------------------------------------------------------- |
+| [Sphinx autodoc](https://www.sphinx-doc.org/en/stable/usage/extensions/autodoc.html) | Python     | Imports the modules it documents and warns that their import-time side effects run; the documented code is trusted. |
+| [TypeDoc](https://github.com/TypeStrong/typedoc/security/policy)                     | TypeScript | Documentation generator; patches the latest minor version and accepts community fixes for the one before.           |
+| [Hugo](https://gohugo.io/about/security/)                                            | Go         | Templates and configuration are trusted, content is not, and output is escaped by default.                          |
+
+## Trusted
+
+An attacker who controls any of these already has what sphinx-argparse-cli could give them:
+
+- The Sphinx configuration and the documents that use the directive.
+- The documented project, its modules and the parser factory, which run during the build.
+
+## Untrusted inputs
+
+| Input             | Comes from                             | What sphinx-argparse-cli guarantees                            |
+| ----------------- | -------------------------------------- | -------------------------------------------------------------- |
+| Directive options | The `.rst` source of the documentation | Select only which module, function and parser parts to render. |
+
+## Threats and controls
+
+| STRIDE                 | Threat                                               | Control                                                                                                              |
+| ---------------------- | ---------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------- |
+| Tampering              | A help string with `.. raw:: html` reaches the HTML. | By design: help text is reStructuredText from trusted code; disable `raw` in Sphinx if contributors are not trusted. |
+| Elevation of privilege | Building docs runs the project's code.               | Accepted, as in Sphinx autodoc.                                                                                      |
+
+## Release chain
+
+Releases run on the shared [tox-dev/github-workflows](https://github.com/tox-dev/github-workflows) pipeline. A release
+App, not a person, pushes the release commit and tag; the build installs its tools from a hash-checked `pylock.toml`
+resolved with a seven-day upload cooldown by default; PyPI receives the files through
+[trusted publishing](https://docs.pypi.org/trusted-publishers/) with attestations, and the GitHub Release is immutable.
+Every workflow pins third-party actions to a commit SHA.
+
+## Vulnerability or bug
+
+| Report                                                 | Verdict             | Why                                    |
+| ------------------------------------------------------ | ------------------- | -------------------------------------- |
+| A `:module:` of `a.b` imports `c` instead              | Vulnerability       | Import target is a guarantee.          |
+| A help string with `.. raw:: html` adds a script       | Not a vulnerability | Help text is trusted reStructuredText. |
+| The parser factory deletes files during the docs build | Not a vulnerability | Documented code is trusted.            |
+| Subcommand order differs from `--help`                 | Bug                 | Rendering only.                        |
+
+## Severity
+
+- **High:** the directive imports a module or calls a function other than the ones its options name.
+- **Medium:** the extension reads or writes files beyond the Sphinx build's own outputs.
+- **Low:** a crash on an unusual parser instead of a clean build error.
+
+## Accepted risks
+
+- Building documentation executes the documented module's import-time code and parser factory.
+
+## Keeping this current
+
+Revisit this model when sphinx-argparse-cli gains a new input source, network access, a new way to run code, or a new
+place it writes to, and whenever an advisory is published.