[waleedlatif1]

Summary

• The copilot permission gate decided by catalog lookup alone. sim_cli has no catalog entry, so no CLI command was ever held for approval, and the CLI's own confirmation is answered by --yes in model-supplied argv
• The worker exposes deploy, promote, run, cancel, and MCP deploy/delete only through sim_cli, so the requiresApproval flags on promote_to_live, deploy_as_*, run_workflow, cancel_workflow_run, delete_workspace_mcp_server, and call_integration_tool guarded nothing
• Added resolveEmbeddedCliCommand to sim/embed: resolves argv with the CLI's real commander tree and parser (stopped at preAction, nothing runs) to the dispatched command's path and the v2 operations it declares. connect/OperationClient already refuse undeclared operations, so this bounds what the call can do
• cliRequestRequiresApproval holds a CLI call when any declared operation is the capability of an approval-gated catalog tool (activate + rollback, deploy/undeploy, chat publish/unpublish, MCP tool deploy/undeploy, MCP server delete, execute/resume, cancel run, tool execute). Read commands stay ungated. Argv the parser cannot follow is held
• Both lanes ask the same predicate: the dispatch gate (by execName) and the in-band route (with params)
• Always-allow is still keyed by the row's display name, but covers a CLI call only when that name matches the command the parser resolves, so a row named for another command (or generic sim_cli) never carries consent across commands

Type of Change

• Bug fix

Testing

• New handler tests run the real router + CLI parser through prePersistClientExecutableToolCall: a workflows activate create … --yes frame is stamped awaiting_approval; always-allow on a sibling command or sim_cli does not cover it
• Router lane rows for activate, rollback (gated) and list (ungated)
• resolveEmbeddedCliCommand tests: global flags before/between path segments, aliases; help/unknown/usage-error/--dry-run argv resolve to no command and running each through runEmbeddedCli makes zero HTTP calls
• Reverted each guard and confirmed its test goes red (catalog-only gate, display-name always-allow, missing rollback mapping)
• handlers.test.ts, router.test.ts, permission.test.ts, embed.test.ts, lint, sim-cli type-check, and the comment-hygiene/file-names/utils/explicit-any/boundaries audits locally; full suite, check:audits and type-check in CI

Checklist

• Code follows project style guidelines
• Self-reviewed my changes
• Tests added/updated and passing (new tests pass the test-audit authoring gate)
• No new warnings introduced
• I confirm that I have read and agree to the terms outlined in the Contributor License Agreement (CLA)

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

1 Skipped Deployment

Project	Deployment	Actions	Updated
docs	Skipped		Nov 2, 2026 9:21pm UTC

[cubic-dev-ai]

No issues found across 9 files

Confidence score: 5/5

• Automated review surfaced no issues in the provided summaries.
• No files require special attention.

_{Turn on auto-fix | Re-trigger cubic}

[greptile-apps]

Confidence Score: 5/5

[High impact] The PR appears safe to merge; no actionable issues were found.

Summary

Makes Copilot approval checks follow the command that sim_cli would actually run.

• Copilot approval follows the capability a CLI command can use.

No actionable issues were found. Review was based on source inspection; tests were not run.

Diagram

%%{init: {'theme': 'neutral'}}%%
flowchart TD
  A[sim_cli request] --> B[Validate request]
  B --> C[Resolve command before its action]
  C --> D{Declared operation needs approval?}
  D -->|No| E[Normal execution]
  D -->|Yes| F{Execution lane}
  F -->|In-band| G[Refuse and request checkpoint lane]
  F -->|Checkpoint| H{Saved consent matches command?}
  H -->|Yes| E
  H -->|No| I[Wait for user approval]
  I -->|Allow| E
  I -->|Skip| J[Do not execute]

Loading

_{Reviews (3) · Last reviewed commit: "fix(copilot): type the sim_cli test fram..." · Reviewed by Greptile}

[waleedlatif1]

@greptile

[waleedlatif1]

@cubic-dev-ai review this PR

[cubic-dev-ai]

@cubic-dev-ai review this PR

@waleedlatif1 I have started the AI code review. It will take a few minutes to complete.

[cubic-dev-ai]

No issues found across 9 files

Confidence score: 5/5

• Automated review surfaced no issues in the provided summaries.
• No files require special attention.

_{Turn on auto-fix | Re-trigger cubic}

[waleedlatif1]

@greptile

[waleedlatif1]

@cubic-dev-ai review this PR

[cubic-dev-ai]

@cubic-dev-ai review this PR

@waleedlatif1 I have started the AI code review. It will take a few minutes to complete.

[cubic-dev-ai]

No issues found across 9 files

Confidence score: 5/5

• Automated review surfaced no issues in the provided summaries.
• No files require special attention.

_{Turn on auto-fix | Re-trigger cubic}