feat(mcp,cdp): lint_page_html tool with static HTML parser (#79) - #222
Merged
Conversation
Adds `plumb_cdp::snapshot_from_html(html, base_url)` and the matching caps `SNAPSHOT_FROM_HTML_INPUT_BYTE_CAP` (1 MiB) and `SNAPSHOT_FROM_HTML_ELEMENT_CAP` (10 000), plus two new `CdpError` variants (`HtmlInputTooLarge`, `HtmlElementLimitExceeded`). The function builds a `PlumbSnapshot` from a static HTML string with no Chromium launch, no JavaScript execution, and no resource fetching — a pure parser pass over `scraper::Html`. `computed_styles` is empty and `rect` is `None` because there is no rendering pipeline available from a static document; the rest of the snapshot mirrors what `PlumbSnapshot::canned` produces. Unit tests cover the canonical html/head/body shape, byte-deterministic output across two calls, the byte-cap refusal, and the element-cap refusal. Refs #79. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Adds the `lint_page_html` MCP tool — `{html, base_url}` in, the same
MCP-compact response shape as `lint_url` out, no Chromium launch. The
handler delegates to `plumb_cdp::snapshot_from_html`, runs the rule
engine, and returns through `build_lint_url_result` so token-budget
guarantees match `lint_url` (compact-only; the static-HTML path has no
rendered styles or geometry that the full envelope could enrich).
Empty `html`, empty `base_url`, and oversized inputs (above the 1 MiB
byte cap or 10 000 element cap enforced by `snapshot_from_html`) all
surface as JSON-RPC `invalid_params` (-32602) so a chatty agent can
react without parsing the underlying `CdpError`.
Protocol tests cover the happy path (compact payload shape), the
empty-`html` invalid-params case, and the byte-cap invalid-params case.
Refs #79.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Extends the existing `mcp_initialize_and_tools_list` end-to-end stdio test to assert that `lint_page_html` shows up in `tools/list`, that its `inputSchema` declares `html` and `base_url` as string properties, and that both fields are listed under `required`. Also adds the tool to the user-facing tool table in `docs/src/mcp.md`. Refs #79. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
…page-html # Conflicts: # crates/plumb-cdp/Cargo.toml # crates/plumb-mcp/src/lib.rs # crates/plumb-mcp/tests/mcp_protocol.rs
aram-devdocs
added a commit
that referenced
this pull request
May 6, 2026
9 tasks
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Closes #79.
Summary
plumb_cdp::snapshot_from_html(html, base_url)— a pure parser pass overscraper::Htmlthat materializes aPlumbSnapshotwithout launching Chromium. Capped at 1 MiB of input and 10 000 elements; oversized inputs return typedCdpError::HtmlInputTooLarge/HtmlElementLimitExceeded.lint_page_htmlMCP tool ({ html, base_url }in, MCP-compact response shape out). Empty arguments and oversized inputs surface as JSON-RPCinvalid_params(-32602).docs/src/mcp.mdtool table to list the new tool.The static-HTML path has no rendering, so
computed_stylesis empty andrectisNone; the tool catches structural rules but not rendering-dependent ones — documented in the tool description and indocs/src/mcp.md.Test plan
cargo test -p plumb-cdp --lib snapshot_from_html— 4 unit tests (canonical shape, byte-deterministic, byte-cap refusal, element-cap refusal).cargo test -p plumb-mcp --test mcp_protocol lint_page_html— 3 protocol tests (happy path, empty html, byte-cap).cargo test -p plumb-cli --test mcp_stdio mcp_initialize_and_tools_list— stdio test assertslint_page_htmlintools/listwith the expectedinputSchema.cargo clippy --workspace --all-targets— zero warnings.cargo deny check— advisories / bans / licenses / sources all OK.🤖 Generated with Claude Code