Files
PaperClipAI/docs
Eldar PolitkinandClaude Opus 5.5 b298f5ce1d docs: allow the chat webhook ingress hostname in private mode (#18301)
## Thinking Path

> - Paperclip is the open source app people use to manage AI agents for
work
> - Native chat connectors receive provider webhooks at
`/api/chat-webhooks/<publicId>/<provider>`
> - `PAPERCLIP_CHAT_WEBHOOK_PUBLIC_URL` lets the webhook ingress use a
different host than the board
> - In `authenticated` + `private` mode, the private-hostname guard runs
on every request, also on the webhook route
> - The docs say that the ingress URL is not a trusted host. They do not
say that the operator must allow the ingress hostname
> - So a private instance behind a webhook-only tunnel answers 403 to
every provider callback, and the operator gets no hint in the docs
> - This pull request adds that step to the
`PAPERCLIP_CHAT_WEBHOOK_PUBLIC_URL` row
> - The benefit is that operators of private instances can connect Teams
(or another webhook provider) without reading the guard source

## Linked Issues or Issue Description

Refs #13468 (the next failure on the same path, after this step is
done). Refs #15185.

No issue exists for this docs gap. I searched open and closed issues and
PRs for "hostname is not allowed webhook",
"PAPERCLIP_CHAT_WEBHOOK_PUBLIC_URL", "private hostname guard webhook",
"allowed-hostname tunnel 403" and "chat webhook 403". Only #14492
mentions the variable, and it is about Telegram long polling.

**What happened?**
Microsoft Teams callbacks through a webhook-only tunnel got `403
{"error":"This hostname is not allowed for this Paperclip instance.
..."}`. The request did not reach the Teams route. A `Host` rewrite at
the tunnel did not help, because the guard reads `X-Forwarded-Host`
first.

**What did you expect?**
The docs for `PAPERCLIP_CHAT_WEBHOOK_PUBLIC_URL` tell the operator to
add the ingress hostname to `PAPERCLIP_ALLOWED_HOSTNAMES` in private
mode.

**Steps to reproduce**
1. Run `authenticated` mode with `private` exposure.
2. Set `PAPERCLIP_CHAT_WEBHOOK_PUBLIC_URL=https://<ingress-host>`. Do
not add `<ingress-host>` to `PAPERCLIP_ALLOWED_HOSTNAMES`.
3. Send `POST
https://<ingress-host>/api/chat-webhooks/<publicId>/microsoft-teams`
through the tunnel.
4. The response is 403 from the private-hostname guard.

**Paperclip version**
v2026.1023.0 (Docker image), and the same code on `master`.

**Deployment mode**
`authenticated` + `private`, Docker, Postgres, Cloudflare tunnel that
forwards only the Teams webhook path.

## What Changed

- `docs/deploy/environment-variables.md`: the
`PAPERCLIP_CHAT_WEBHOOK_PUBLIC_URL` row now tells private-mode operators
to add the ingress hostname to `PAPERCLIP_ALLOWED_HOSTNAMES`. It also
says that the guard reads `X-Forwarded-Host` before `Host`.

## Verification

- Read the guard: `server/src/middleware/private-hostname-guard.ts`
reads `x-forwarded-host` first, then `host`. `server/src/app.ts` mounts
the guard for all routes, also `/api/chat-webhooks`.
- Live check on our instance (v2026.1023.0, Microsoft 365 commercial
tenant, Cloudflare tunnel): before the allowlist entry, `POST
/api/chat-webhooks/<publicId>/microsoft-teams` returned 403 with the
guard message. After we added the ingress hostname to
`PAPERCLIP_ALLOWED_HOSTNAMES`, the same request reached the Teams route
(`404 {"error":"Chat endpoint not found"}` for an unknown publicId).
- `node scripts/check-forbidden-tokens.mjs`: `No forbidden tokens
found.`
- Docs-only change. No code or test change.

## Risks

- Low risk. Docs only. No behavior change.

> For core feature work, check [`ROADMAP.md`](ROADMAP.md) first and
discuss it in `#dev` before opening the PR. Feature PRs that overlap
with planned core work may need to be redirected - check the roadmap
first. See `CONTRIBUTING.md`.

## Model Used

- Anthropic Claude, model ID `claude-opus-5-5` (Opus 5.5), in Claude
Code. Tool use (shell, GitHub CLI) and code execution. Default reasoning
mode. A human operator owns the Paperclip instance and the evidence.

## Checklist

- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either (a) linked existing issues with `Fixes: #` / `Closes
#` / `Refs #` OR (b) described the issue in-PR following the relevant
issue template
- [x] I have not referenced internal/instance-local Paperclip issues or
links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip`
URLs)
- [x] My branch name describes the change (e.g. `docs/...`, `fix/...`)
and contains no internal Paperclip ticket id or instance-derived details
- [x] I have run tests locally and they pass
- [ ] I have added or updated tests where applicable
- [x] I have updated relevant documentation to reflect my changes
- [x] I have considered and documented any risks above
- [ ] All Paperclip CI gates are green
- [ ] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
- [x] I will address all Greptile and reviewer comments before
requesting merge

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-16 14:52:44 -06:00
..