build(deps): bump dompurify from 3.2.6 to 3.4.15 in /antora

[dependabot]

Bumps dompurify from 3.2.6 to 3.4.15.

Release notes

Sourced from dompurify's releases.

DOMPurify 3.4.15

• Added better clobbering hardening when XML content is involved, thanks @​gnyselcuk
• Added several smaller hardening and edge-case improvements, thanks @​leechristensen
• Bumped several dependencies where possible

DOMPurify 3.4.14

• Fixed an issue with possible bypasses when risky tags are allow-listed, thanks @​AlirezaRouhbakhsh
• Fixed a couple of edge cases with mixed document contexts, thanks @​fishjojo1
• Added the SVG pointer-events and vector-effect presentation attributes to the allow-list, thanks @​Jaybhade
• Conducted another refactoring run, removed dead branches and duplicated logic, flattened attribute validation
• Updated the documentation in several spots, README, wiki, etc., thanks @​Akokonunes
• Updated several development dependencies and CI workflow actions

DOMPurify 3.4.13

• Fixed an issue with hook removal during IN_PLACE sanitization, thanks @​koyokr
• Fixed an issue with hooks potentially bypassing the clone guard, thanks @​AkshayjainG
• Fixed an issue with DOM clobbering via ownerDocument during IN_PLACE, thanks @​AkshayjainG
• Bumped several dependencies where possible

DOMPurify 3.4.12

• Fixed an issue where a hook would not get called for custom elements, thanks @​Rikuxx0
• Hardened the handling of hooks removing elements, @​mkrause-bee360
• Added support for a few new SVG attributes, thanks @​cbn-falias & @​Develop-KIM
• Hardened the handling of declarative partial updates
• Updated the documentation is several spots, README, wiki, etc.
• Bumped several dependencies where possible

DOMPurify 3.4.11

• Fixed an issue with a leaky config for hooks via setConfig, thanks @​trace37labs
• Bumped vulnerable development dependencies to arrive at plain 0 with npm audit
• Updated the osv-scanner suppression list as no vulnerable dependencies are left for now
• Updated up the linting tool-chain and removed now-redundant lint directives
• Updated the documentation is several spots, README, wiki, etc.
• Bumped several dependencies where possible

DOMPurify 3.4.10

• Refactored codebase for clarity: extracted the public type declarations into types.ts
• Decomposed the three largest sanitizer functions into focused helpers
• Removed duplicated defaults and dead branches, consolidated SAFE_FOR_TEMPLATES scrubbing into single shared path
• Improved per-node performance by hoisting the mXSS probe regexes and testing textContent before innerHTML
• Added a deterministic micro-benchmark harness (npm run bench) with a --compare mode
• Reduced CI cost by running the full three-engine browser suite once per PR
• Refreshed the demos/ folder so every demo runs again, and added a SVG-via-<img> demo
• Documented the bench and test:happydom scripts in the README
• Completed the Attack Classes & Bypass History wiki page
• Bumped several dependencies where possible

DOMPurify 3.4.9

• Further improved the handling of Trusted Types config options, thanks @​offset
• Further improved the handling of IN_PLACE sanitization, thanks @​mozfreddyb

... (truncated)

Commits

• 1d7460c release: 3.4.15 (#1609)
• 4e6fe24 release: 3.4.14 (#1587)
• 3067f77 release: 3.4.13 (#1562)
• a9ca1e5 release: 3.4.12 (#1537)
• 0cae518 release: 3.4.11 (#1494)
• 6ee5716 release: 3.4.10 (#1478)
• 5210247 release: 3.4.9 (#1459)
• bcdd828 release: 3.4.8 (#1439)
• ca30f07 release: 3.4.7 (#1414)
• bb7739e release: 3.4.6 (#1394)
• Additional commits viewable in compare view

Install script changes

This version adds prepare script that runs during installation. Review the package contents before updating.

Dependabot will resolve any conflicts with this PR as long as you don't alter it yourself. You can also trigger a rebase manually by commenting @dependabot rebase.

---

Dependabot commands and options

You can trigger Dependabot actions by commenting on this PR:

• @dependabot rebase will rebase this PR
• @dependabot recreate will recreate this PR, overwriting any edits that have been made to it
• @dependabot show <dependency name> ignore conditions will show all of the ignore conditions of the specified dependency
• @dependabot ignore this major version will close this PR and stop Dependabot creating any more for this major version (unless you reopen the PR or upgrade to it yourself)
• @dependabot ignore this minor version will close this PR and stop Dependabot creating any more for this minor version (unless you reopen the PR or upgrade to it yourself)
• @dependabot ignore this dependency will close this PR and stop Dependabot creating any more for this dependency (unless you reopen the PR or upgrade to it yourself)
  You can disable automated security fix PRs for this repo from the Security Alerts page.

[coderabbitai]

Important

Review skipped

Bot user detected.

To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Organization UI

Review profile: CHILL

Plan: Advanced

Run ID: 08f351be-afa0-4f4e-894e-60552183cf28

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

---

_{Comment @coderabbitai help to get the list of available commands.}

[github-actions]

🚀 Preview is available at https://331d5fe3.enterprise-contract.pages.dev

[fullsend-ai-review]

🤖 Finished Review · ✅ Success · Started 8:00 PM UTC · Completed 8:10 PM UTC

Commit: cd01138 · View workflow run →

Runtime: claude · Model: opus → claude-opus-4-7 · Effort: high · Cost: $3.83

[fullsend-ai-review]

Looks good to me

[dependabot]

OK, I won't notify you again about this release, but will get in touch when a new version is available. If you'd rather skip all updates until the next major or minor version, let me know by commenting @dependabot ignore this major version or @dependabot ignore this minor version.

If you change your mind, just re-open this PR and I'll resolve any conflicts on it.

[fullsend-ai-retro]

🤖 Finished Retro · ✅ Success · Started 3:13 PM UTC · Completed 3:19 PM UTC

Commit: cd01138 · View workflow run →

Runtime: claude · Model: opus → claude-opus-4-7 · Effort: high · Cost: $2.01

[fullsend-ai-retro]

Retro on PR #558 — Dependabot bump dompurify 3.2.6 → 3.4.15

Outcome: Review agent approved a lockfile-only bot PR in ~11 min for $3.83 (opus, high effort). Human closed the PR without merging the next day. The repo actually uses Renovate, not Dependabot, and dompurify is a transitive dep of mermaid — no direct import in the tree.

No new proposals filed. Every improvement this run surfaces is already tracked by an open issue upstream; per the retro skill's guidance, evidence goes here instead of duplicate issues.

Evidence for existing issues

• fullsend-ai/agents#257 — Pre-review early exit for bot-authored dependency-update PRs. This run is a textbook case: single-file lockfile diff (+3/-7) from Dependabot, size: XS, no source-code reference to the bumped package. The orchestrator did classify it "trivial" (≤5 tool-call cap on sub-agents) but still spent $3.83 and 11 min. An earlier bot-exit would have saved essentially all of it.
• fullsend-ai/fullsend#6666 — Stale sonnet mapping inflates sub-agent cost. The risk-assessment and style-conventions sub-agents both tried claude-sonnet-4-5@20250929, which is unavailable on this Vertex deployment; each retried 3× before giving up (~2 min wasted, non-trivial share of the wall-clock and cost). Two of four sub-agent verdicts were therefore missing when the approval was issued.
• fullsend-ai/agents#497 & fullsend-ai/fullsend#5777 — Route model/effort by scope, not just author. Opus + high effort on a lockfile-only transitive bump is over-provisioned; scope-based downgrade would have caught this case whether the author was a bot or a human.
• fullsend-ai/fullsend#6991 / #6968 — Skip redundant re-review. Two full agent iterations ran within the single dispatch (Agent runs: 2): iteration 1 (4m03s, $2.26) already reached "approve"; iteration 2 (3m31s, $1.57) re-read the same skill files, re-fetched the PR, and re-dispatched the same four sub-agents. Roughly 40% of the run's cost was duplicative. The existing issues target cross-run duplication, but the same underlying signal (unchanged input → unchanged output) applies intra-run and could be gated by trace-fitness on iteration 1.
• fullsend-ai/agents#189 & #980 — worth keeping in view as counter-cases if any of the above are implemented: the review still needs to catch downgrades, malicious install-time scripts (dompurify 3.4.x added a prepare script — the security sub-agent did flag this correctly), and unaudited submodule/lockfile substance changes.

One repo-side observation, not filed

The repo has no .github/dependabot.yml and uses Renovate exclusively; this Dependabot PR appears to be a GitHub-Security-Alerts–initiated PR that the maintainer chose to close (they use Renovate for the same bumps). Not a fullsend problem — noting only for context on why the review effort was ultimately wasted.

Discovery notes

• Agents repo: fullsend-ai/agents@v0.41.0 (SHA 10560252b0ad).
• Model resolved: opus → claude-opus-4-7 (v2.1.258).
• Trace artifact: fullsend-review.zip.