[avrabe]

Closes #1105

Summary

test_scanner::detect_language had no entry for Robot Framework (.robot, .resource), so # rivet: verifies REQ-NNN markers in Robot suites were silently skipped. Reported downstream: 145 inflated warnings from a project whose tests are Robot Framework suites.

Same shape as the shell gap (REQ-319), the yaml/toml gap (REQ-352) and the lean/rocq gap (#992) — the pattern itself was never the obstacle; scan_file bailed before any pattern ran.

Fix

• Add robot as its own language in default_patterns, using # comments.
• Map .robot and .resource to robot in detect_language.
• Add robot to find_enclosing_function's early-return list (Robot test cases are names under *** Test Cases ***, not functions — guessing a name would be the REQ-326 misattribution shape the Lean/Rocq guard already prevents).

Acceptance criteria from the issue

• .robot / .resource files are scanned with # comments. detect_language now maps both extensions to "robot", and default_patterns carries a matching entry.
• A marker placed on the line before a test case name is credited. robot_framework_files_are_scanned_for_markers places the marker on the line above My Test Case (and My Keyword in a .resource file) and asserts both ids are found.
• A test with a .robot fixture asserts the marker is credited. Added robot_framework_files_are_scanned_for_markers in rivet-core/src/test_scanner.rs.
• A negative control drops the extension mapping and watches it fail. robot_markers_fall_back_to_file_and_line plants def helper( above the marker; removing the "robot" guard from find_enclosing_function latches the generic catch-all onto helper and the assert_ne!(test_name, "helper") fails. The two assertions together kill a mutant that drops either the extension mapping or the enclosing-function guard.

Test plan

• cargo test -p rivet-core --lib test_scanner:: — 27 passed (both new tests included).
• CI green.

---

Generated by Claude Code

[avrabe]

Security Audit (RustSec) failing — standing down on this one as it is not this PR's.

The diff is a single Rust-logic change in rivet-core/src/test_scanner.rs; no Cargo.toml / Cargo.lock / deny.toml edits. The job reports six advisories:

• RUSTSEC-2024-0384 (instant unmaintained) — already in deny.toml ignore list
• RUSTSEC-2026-0251 (sized-chunks unmaintained) — already in deny.toml
• RUSTSEC-2023-0126 (im unsound, 2023-02-04)
• RUSTSEC-2026-0308 (salsa use-after-free) — already assessed not-applicable in deny.toml
• RUSTSEC-2026-0205 (scc double-free, 2026-07-06)
• RUSTSEC-2026-0255 (sized-chunks panic-safety unsound, 2026-08-11)

All six predate this PR, and the same check was green on the latest scheduled main run (2026-10-19). The audit job in .github/workflows/ci.yml:706 installs cargo-audit unpinned (cargo install cargo-audit --locked), so a release landing between that run and today likely tightened its default (treating unmaintained/unsound as errors rather than warnings) — the --ignore list in the workflow covers wasmtime-era advisories but none of the six now surfacing. The three already in deny.toml are not duplicated to the workflow's --ignore list because they historically only mattered to cargo deny.

No fix PR exists yet. The right fix is a separate CI-only change: either pin cargo-audit to a version matching last run's defaults, or extend the workflow's --ignore list to mirror deny.toml. Not porting it into this PR to keep the diff focused on #1105.

Will spend my one allotted re-run to confirm it reproduces, then wait on the fix landing on main.

---

Generated by Claude Code

[avrabe]

Playwright E2E failing — also not this PR's.

BEWARE: your OS is not officially supported by Playwright; installing dependencies for ubuntu26.04-x64 as a fallback.
Cannot install dependencies for ubuntu26.04-x64 with Playwright 1.58.2!
Error: ERROR: Playwright does not support chromium on ubuntu26.04-x64

GitHub's ubuntu-latest moved to ubuntu-26.04 and the pinned Playwright 1.58.2 doesn't recognize it. The diff touches no tests/playwright/** and no workflow file. Same job was green on the latest scheduled main run (2026-10-19), so the runner image bumped between then and today.

Fix for a separate CI PR: either bump the Playwright pin to a version that supports 26.04, or pin the runner to ubuntu-24.04. Not porting into this PR to keep it focused on #1105.

---

Generated by Claude Code

[github-actions]

📐 Rivet artifact delta

No artifact changes in this PR. Code-only changes (renderer, CLI wiring, tests) don't touch the artifact graph.

[avrabe]

Cargo Deny failing — also not this PR's.

Two newly-published wasmtime 48.0.6 CVEs:

• RUSTSEC-2026-0378 — panic lowering a host-created component-model stream (announced 2026-10-20)
• RUSTSEC-2026-0379 — Cranelift aarch64 br_table jumps to incorrect target (announced 2026-10-20)

Both have the same fix: cargo update -p wasmtime to >=48.0.7, <49.0.0 OR >=49.0.4. The diff touches no Cargo.toml / Cargo.lock, and rivet-core only pins the parent wasmtime version (b24baa6 bumped it to 48.0.6, which these advisories flag). The same job was green on the latest scheduled main run (2026-10-19) because the advisories hadn't been published yet. Main is now red against the fresh advisory DB too — unlike the audit/playwright failures above this is a real vulnerability that needs a wasmtime bump on main, not a CI config change.

Not porting into this PR to keep the diff focused on #1105. A separate build(deps): wasmtime 48.0.6 → 48.0.7 PR will clear it, same shape as b24baa6.

---

Generated by Claude Code

[codecov]

Codecov Report

❌ Patch coverage is 92.68293% with 3 lines in your changes missing coverage. Please review.

Files with missing lines	Patch %	Lines
rivet-core/src/test_scanner.rs	92.68%	3 Missing ⚠️

📢 Thoughts on this report? Let us know!

[avrabe]

Superseded by #1108 (merged as cc3b807, ships in v0.41.1, REQ-407): .robot / .resource markers are scanned there too, and a marker is attributed to the Robot test case or keyword it precedes or sits in (falling back to file:line outside any case), with a negative control. This branch now conflicts with main and fails on advisories cleared in #1108.