plans/README.md
Generated by the improve skill on 2026-07-10, planned against commit
c63cb120 (Slidev v52.16.0). Each plan is a self-contained handoff for an
executor with zero prior context: read the plan fully before starting, run its
verification commands, honor its STOP conditions, and update your row below when
done.
Advisor constraint: these plans were produced read-only. No source code was
modified. The audit that produced them is summarized in each plan's "Why this
matters" / "Current state". Verification baseline for the whole repo:
pnpm install && pnpm build && pnpm typecheck && pnpm lint && pnpm test
(tests require a prior pnpm build; plan 003 adds a verify aggregate).
Ordered by leverage and dependencies: cheap/low-risk wins first, then high-confidence security + correctness, then larger refactors. You may cherry-pick — only the edges in "Dependency notes" are hard requirements.
| Plan | Title | Priority | Effort | Risk | Depends on | Status |
|---|---|---|---|---|---|---|
| 001 | CI type-checks on every PR | P1 | S | LOW | — | DONE (plan file removed) |
| 002 | Cache the pnpm store in CI | P2 | S | LOW | — | DONE (plan file removed) |
| 003 | Aggregate verify script + document build-before-test | P2 | S | LOW | — | DONE (plan file removed) |
| 004 | Fix catalog drift (runtime deps on catalog:dev) | P2 | S | LOW | — | DONE (plan file removed) |
| 005 | Add contributor AGENTS.md | P3 | S | LOW | — | TODO |
| 006 | Replace exec shell-string with execFile (edit shortcut) | P2 | S | LOW | — | DONE (plan file removed) |
| 007 | Guarantee Chromium teardown on export failure | P1 | S | LOW | — | DONE (plan file removed) |
| 008 | Guard against circular src: slide imports | P1 | S | LOW | — | DONE (plan file removed) |
| 009 | parseRangeString lower-bound / NaN validation | P2 | S | LOW | — | TODO |
| 010 | Harden getSlidePath against unknown slide | P2 | S | LOW-MED | — | TODO |
| 011 | 404 on out-of-range slide-patch request | P3 | S | LOW | — | TODO |
| 012 | Fix no-op HMR utils refresh (missing await) | P2 | S | LOW | — | TODO |
| 013 | Free port + cleanup for build's temp servers | P2 | S-M | MED | — | TODO |
| 014 | Confine deck-controlled file reads (snippets + src:) | P1 | M | MED | — | DONE (plan file removed) |
| 015 | Validate paths in dev-server write sinks | P1 | S-M | MED | — | DONE (plan file removed) |
| 016 | Confine export output path from deck exportFilename | P2 | S | LOW | — | TODO |
| 017 | Validate WebSocket Origin on privileged ws handlers | P2 | M | MED | — | TODO |
| 018 | Enforce --remote auth server-side (not just client) | P3 | M-L | MED | — | TODO |
| 019 | Key getRoots() cache by entry (multi-entry build/export) | P2 | M | MED | — | TODO |
| 020 | O(1) slide lookup Map (fix O(n²) TOC) | P2 | S | LOW | — | TODO |
| 021 | Consolidate divergent addToTree TOC builders | P2 | M | MED | — | TODO |
| 022 | Test the export pipeline (characterization tests) | P1 | M | LOW | — | TODO |
| 023 | Decompose god functions in export.ts/cli.ts | P3 | L | MED | 022 | TODO |
| 024 | Incremental HMR parse cache | P3 | L | MED | — | TODO |
| 025 | Guard skills/-vs-docs/ drift in CI | P3 | S-M | LOW | — | TODO |
Status values: TODO | IN PROGRESS | DONE | BLOCKED (one-line reason) |
REJECTED (one-line rationale).
If you don't run them all, start with the low-risk, high-confidence set (all clean verification stories): 001, 002, 003, 004, 006, 007, 008, 014, 015 — plus 022 early, because it unblocks the export refactor and de-risks 007/016/023.
exportSlides/serve handler until the
export characterization tests exist (023's only safety net; there are no serve
tests, so its Step 3 is manual + STOP-gated).addToTree into @slidev/parser as buildTocTree;
022 tests the export TOC path. If 021 lands first, 022 tests the shared builder;
if 022 lands first, 021 keeps its test green. Land either order, but re-run the
other's tests.utils refresh actually run (awaited);
024 reduces the parse cost that makes such per-HMR work expensive. If 012's
cost proves high before 024 lands, see 012's STOP condition.packages/client/logic/slides.ts (getSlide/
getSlidePath). Whichever lands second must preserve the other's change
(010 = undefined guard, 020 = lookup Map). Small merge; not a hard edge.014, 015, 016, 017, 018 all harden the dev-server + parser trust boundary that
Slidev's existing vite/importGuard.ts began. 014 (arbitrary file read via
snippets/src:) and 015 (traversal writes) are the highest-confidence,
most-actionable; 018 is the larger architectural lift (real server-side auth).
Frame all of them as defensive maintenance for the "untrusted deck" and
"exposed dev server" threat models — consistent with the maintainers' own
hardening intent.
export.ts:630-655 "empty catches" — NOT a bug: a deliberate 4-location
Playwright-resolution fallback chain ending in a helpful throw.importGuard.ts skipping dynamic imports with ${ template specifiers —
accepted limitation; defense-in-depth atop Vite's server.fs.strict, does not
itself widen fs.allow.resolver.ts catch {} blocks (:116/148/157/162) — deliberate best-effort
resolution fallbacks; not load-bearing error swallowing.taze.config.ts:10).plantuml.com (parser/src/config.ts:37) —
by-design/documented; a docs note at most, not a code change.DEPS-02 minor version drift (@types/katex 0.16 vs katex 0.17; the
@hedgedoc patch; @lillallol/outline-pdf bus-factor) — low leverage; monitor,
no dedicated plan.smoke.yml — a documented pnpm-v11
limitation, not a defect.Surfaced during the audit but out of the "fix" set; each is grounded in repo evidence and would be a design/spike plan if pursued:
packages/vscode/src/lmTools.ts, over @slidev/parser's existing stringify/
parseSlide).skills/ generator (scripts/gen-skills.ts) — the automation
behind plan 025's drift guard.lmTools.ts into CLI
subcommands / an MCP server).integrations/snapshots.ts:25 TODO).