id(web): inline HTML (///…) is backslash-escaped on save — tags become literal text

Problem

In the id web editor, any inline HTML in a markdown file other than <img> / <a><img></a> is destroyed on save: the tags come back backslash-escaped as literal text.

Verified probe on main a2eef6f7 (standalone crate compiling the verbatim pkgs/id/src/web/markdown.rs codec from main; full log in the goal's hidden_files probe-2026-10-25-2312/probe2.log):

Input: water is H2O and x2
- markdown_to_prosemirror produces plain text nodes "", "2", "", …
- Save (prosemirror_to_markdown) emits: water is H\<sub\>2\</sub\>O and x\<sup\>2\</sup\>
- The rendered document now literally shows \<sub\> instead of subscript markup.
Input: some underlined text → save emits some \<u\>underlined\</u\> text.

Root cause

pkgs/id/src/web/markdown.rs:354 — the NodeValue::HtmlInline arm of convert_node special-cases only img/linked-img tags; every other inline HTML tag falls through to text_node_with_marks(html, active_marks), turning the tag into a plain text node. On save, comrak's CommonMark renderer escapes the </> of text nodes, so the tag can never come back as HTML. Note commonmark_options() (markdown.rs:67) already sets render.unsafe = true for the sized-image HTML forms, so raw inline HTML is not rejected by policy elsewhere in this codec.

Why it matters

, , , <kbd>, , <details> etc. are common in notes and technical docs. Opening such a file in the web editor and saving any unrelated edit silently corrupts every inline HTML tag in the file — the same silent-save-corruption class as #872/#873/#880/#881/#882.

Acceptance criteria

A markdown file containing inline HTML (at minimum , , ) round-trips through open → save without the tags being escaped: the saved markdown still contains the tags as HTML (or an explicitly equivalent representation that renders identically).
The existing img / linked-img HTML handling (id(web): parse_img_tag substring search lets data-src/data-width shadow the real img attributes — save rewrites the image URL #872, id(web): linked images [![alt](src)](href) lose their link in the markdown ↔ ProseMirror codec — save strips the href #880, id(web): sized linked images still lose their href on reload — PR #888's <a><img></a> HTML emission has no parse counterpart (and editor parseDOM drops the wrapping link) #893) is unchanged.
New unit tests in pkgs/id/src/web/markdown.rs cover the round-trip for at least // inline HTML.
If full passthrough is judged unsafe, the narrower fix (a whitelist of benign formatting tags) is implemented instead and the rejected set is named in the PR — silently escaping everything is the bug, not a fix.

Suggested direction

Carry non-img HtmlInline content through the PM doc in a form the save direction can emit back as a comrak HtmlInline node (instead of a Text node), mirroring how sized images already round-trip via NodeValue::HtmlInline in json_to_ast (markdown.rs:~1027). The editor (TS) side may need a matching node/mark handling note — flag it in the PR if the PM representation changes.

CODE-155

🤖 Agent note · filed

Agent: new-issue-creator job
Session / ID: new-issue-creator run, 2026-10-25 23:15 CDT (America/Chicago)
Found via: filed this run
Task: HtmlInline arm (markdown.rs:354) turns non-img inline HTML into plain text nodes, so save escapes the tags ( → \<sub\>). Verified this run with a standalone crate compiling the verbatim codec from main a2eef6f: H2O and underlined both round-trip to backslash-escaped literal text (log: goal hidden_files probe-2026-10-25-2312/probe2.log). Queued for the code pilot queue (triage packet — fix needs a representation choice).
Expectation: effort/m — done looks like inline HTML round-tripping unescaped (or a named benign-tag whitelist), with markdown.rs unit tests.
Links: n/a yet (queue packet enqueued this run)

🤖 Agent note · picked up

Agent: queue worker (repo-worker-code), lane A shift
Session / ID: queue task 1131 (resume of parked claim)
Found via: queue claim
Task: Resume from implementation_complete: PR fix(id-web): html_inline passthrough — inline HTML round-trips unescaped (#943) #1009 (html_inline atom node passthrough, Rust + editor.ts) was parked while CI re-ran after a rustfmt fix. All 28 checks now pass on head a8c01ca — merging it and verifying id(web): inline HTML (///…) is backslash-escaped on save — tags become literal text #943 auto-closes.
Expectation: s — done = PR fix(id-web): html_inline passthrough — inline HTML round-trips unescaped (#943) #1009 merged, id(web): inline HTML (///…) is backslash-escaped on save — tags become literal text #943 closed, task completed.
Links: PR fix(id-web): html_inline passthrough — inline HTML round-trips unescaped (#943) #1009 · branch queue/1131-943-html-inline

fix(id-web): html_inline passthrough — inline HTML round-trips unesca…

🤖 Agent note · done

Agent: queue worker (repo-worker-code), lane A shift
Session / ID: queue task 1131
Found via: queue claim (resume of parked claim)
Task: Merged PR fix(id-web): html_inline passthrough — inline HTML round-trips unescaped (#943) #1009 — html_inline atom-node passthrough so inline HTML (_{/^{///…) round-trips unescaped. All 28 CI checks green on the final head; PR merged (squash, d5a0812) and this issue auto-closed.}}
Expectation: done — task 1131 completed with outcome fixed_merged.
Links: PR fix(id-web): html_inline passthrough — inline HTML round-trips unescaped (#943) #1009

Navigation Menu

Repository navigation

id(web): inline HTML (<sub>/<sup>/<u>/…) is backslash-escaped on save — tags become literal text #943

Problem

Root cause

Why it matters

Acceptance criteria

Suggested direction

27 remaining items

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

id(web): inline HTML (<sub>/<sup>/<u>/…) is backslash-escaped on save — tags become literal text #943

Description

Problem

Root cause

Why it matters

Acceptance criteria

Suggested direction

Activity

developing-today commented on Oct 26, 2026

dezren39 commented on Oct 26, 2026

27 remaining items

dezren39 commented on Oct 26, 2026

dezren39 commented on Oct 26, 2026

Metadata

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Issue actions