feat(code-mode): drop Lab action support#74
Open
jmagar wants to merge 5 commits into
Open
Conversation
📝 WalkthroughWalkthroughThis PR completes the removal of Lab-action support from Code Mode, making it upstream-only. Tool ID parsing now rejects ChangesCode Mode Lab-Action Removal
Implementation Plan Documentation
Estimated code review effort🎯 4 (Complex) | ⏱️ ~60 minutes Possibly related PRs
Poem
🚥 Pre-merge checks | ✅ 4 | ❌ 1❌ Failed checks (1 warning)
✅ Passed checks (4 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing Touches📝 Generate docstrings
🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 2
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In `@crates/lab/src/dispatch/gateway/code_mode.rs`:
- Around line 1160-1180: The test code_search_returns_only_upstream_candidates
currently constructs CodeModeBroker with None so search short-circuits; update
the test to create and seed a real GatewayManager/pool and pass it into
CodeModeBroker::new instead of None (or otherwise ensure the broker is wired to
an upstream manager), call broker.search(...) as before, assert results is
non-empty, and add an assertion that at least one candidate.id
starts_with("upstream::") (while keeping the existing assertion that no id
starts_with("lab::")). Use the existing symbols CodeModeBroker::new,
broker.search, and the test function
code_search_returns_only_upstream_candidates to locate where to inject the
manager/pool setup.
In `@docs/superpowers/plans/2026-05-25-code-mode-v2-drop-lab-actions.md`:
- Line 19: The table row is broken by unescaped pipe characters in the inline
command text referencing CLI subcommands (gateway code search|schema|exec);
update the cell in the markdown so the pipes are escaped (e.g., replace | with
\|) or wrap the entire command text in a code span/backticks without raw |
characters, ensuring the reference to crates/lab/src/cli/gateway.rs and the CLI
subcommands (gateway code search, schema, exec) remains intact and the table
renders correctly.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Organization UI
Review profile: ASSERTIVE
Plan: Pro
Run ID: e0349e2d-3675-4444-9afe-1c9c5f9a851a
📒 Files selected for processing (5)
crates/lab/src/cli/gateway.rscrates/lab/src/dispatch/gateway/code_mode.rscrates/lab/src/main.rscrates/lab/src/mcp/server.rsdocs/superpowers/plans/2026-05-25-code-mode-v2-drop-lab-actions.md
Comment on lines
+1160
to
+1180
| async fn code_search_returns_only_upstream_candidates() { | ||
| let registry = super::ToolRegistry::new(); | ||
| let broker = super::CodeModeBroker::new(®istry, None); | ||
|
|
||
| let err = broker | ||
| .schema( | ||
| "lab::gateway.danger", | ||
| super::CodeModeCaller::Scoped { | ||
| scopes: vec!["lab".to_string()], | ||
| subject: Some("subject-1".to_string()), | ||
| }, | ||
| super::CodeModeSurface::Mcp { | ||
| expose_builtin_services: true, | ||
| allow_destructive_actions: true, | ||
| }, | ||
| let results = broker | ||
| .search( | ||
| "movie.search", | ||
| 10, | ||
| super::CodeModeCaller::TrustedLocal, | ||
| super::CodeModeSurface::Cli, | ||
| ) | ||
| .await | ||
| .expect_err("schema for admin-only action must require admin scope"); | ||
| .expect("search ok"); | ||
|
|
||
| assert_eq!(err.kind(), "forbidden"); | ||
| for candidate in &results { | ||
| assert!( | ||
| !candidate.id.starts_with("lab::"), | ||
| "found lab:: candidate after drop: {}", | ||
| candidate.id | ||
| ); | ||
| } |
There was a problem hiding this comment.
Exercise the upstream search path in this regression test.
Lines 1161-1172 construct CodeModeBroker with None, so search() exits with Ok(Vec::new()) before any candidate filtering runs. This will stay green even if lab:: results leak back in once a GatewayManager is configured. Seed a manager/pool here and assert at least one returned id is upstream::....
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
In `@crates/lab/src/dispatch/gateway/code_mode.rs` around lines 1160 - 1180, The
test code_search_returns_only_upstream_candidates currently constructs
CodeModeBroker with None so search short-circuits; update the test to create and
seed a real GatewayManager/pool and pass it into CodeModeBroker::new instead of
None (or otherwise ensure the broker is wired to an upstream manager), call
broker.search(...) as before, assert results is non-empty, and add an assertion
that at least one candidate.id starts_with("upstream::") (while keeping the
existing assertion that no id starts_with("lab::")). Use the existing symbols
CodeModeBroker::new, broker.search, and the test function
code_search_returns_only_upstream_candidates to locate where to inject the
manager/pool setup.
| |---|---|---| | ||
| | `crates/lab/src/dispatch/gateway/code_mode.rs` | Code Mode broker, ID parsing, runner protocol, schema construction, TS bindings | Heavy delete + edit | | ||
| | `crates/lab/src/mcp/server.rs` | MCP tool registration + dispatch for `code_search`/`code_schema`/`code_execute` | Light edit | | ||
| | `crates/lab/src/cli/gateway.rs` | CLI Code Mode subcommands (`gateway code search|schema|exec`) | Light edit (if any `lab::` branches exist) | |
There was a problem hiding this comment.
Fix broken table row caused by unescaped pipe characters.
The inline command text includes | separators, which creates extra table cells and breaks rendering. Escape the pipes or wrap the whole cell content in backticks without raw | characters.
🧰 Tools
🪛 markdownlint-cli2 (0.22.1)
[warning] 19-19: Table column count
Expected: 3; Actual: 5; Too many cells, extra data will be missing
(MD056, table-column-count)
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
In `@docs/superpowers/plans/2026-05-25-code-mode-v2-drop-lab-actions.md` at line
19, The table row is broken by unescaped pipe characters in the inline command
text referencing CLI subcommands (gateway code search|schema|exec); update the
cell in the markdown so the pipes are escaped (e.g., replace | with \|) or wrap
the entire command text in a code span/backticks without raw | characters,
ensuring the reference to crates/lab/src/cli/gateway.rs and the CLI subcommands
(gateway code search, schema, exec) remains intact and the table renders
correctly.
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.
Summary
lab::<service>.<action>ID support from Code Mode socode_search,code_schema, andcode_executeoperate on proxied upstream MCP tools only.lab::IDs with a structuredunknown_toolerror and points agents totool_executefor Lab actions.expose_builtin_services, and the Code Mode Lab-action authorization path while keeping upstream JSON Schema TypeScript bindings.Bead:
lab-elme3.1Test plan
cargo test -p labby --all-features code_mode -- --nocapturecargo test -p labby --all-features mcp::server -- --nocapturecargo test -p labby --all-features --test code_mode_runner -- --nocapturecargo check --manifest-path crates/lab/Cargo.toml --all-featurescargo fmt --all -- --checkcargo clippy --workspace --all-features -- -D warningscargo test -p labby --lib --all-features -- --test-threads=1cargo test --workspace --all-features -- --test-threads=1labby 0.17.4, verified/healthand/ready, confirmedcode_searchreturns upstream ids only, confirmedcode_schema id=lab::radarr.movie.searchreturnsunknown_toolguidance, and rancode_execute12-readonly plus chained fanout smokes successfully.Summary by cubic
Drops Lab action IDs from Code Mode and removes
code_schema. Code Mode now supports only upstream tools viacode_search(JS-sandbox catalog filtering) andcode_execute; anylab::ID returnsunknown_toolwith a hint to usetool_execute.Refactors
lab::ID parsing; onlyupstream::server::toolis supported.code_schema; docs and CLI help updated forcode_search+code_executeonly.code_searchruns constrained JS over an inlined catalog; 256KB soft cap and 512KB hard cap with stable pruning.CodeModeConfigcaps (max_response_bytes,max_response_tokens) and response envelope trimming.code_mode_timeoutandcode_mode_fuel_exhaustedmap to HTTP 504.code_mode_wasmviajavy+wasmtime; addedscripts/refresh-javy-plugin.sh.Mcp { allow_destructive_actions }; refreshed MCP descriptions.libclang-devfor Code Mode builds, lighter release image,cargo-deny/license updates, test fixes.Migration
tool_executefor Lab actions instead oflab::Code Mode IDs.tool_execute(name="radarr", arguments={action:"movie.search", params:{query:"Matrix"}}).code_searchto filter the upstream catalog, then pass returnedupstream::server::toolIDs tocode_execute. Remove anycode_schemacalls.Written for commit ac5a374. Summary will update on new commits. Review in cubic
Summary by CodeRabbit
Breaking Changes
Documentation