fix(mcp-bridge): harden argument handling for routed tools

[hyperpolymath]

Summary

Harden argument handling in the MCP bridge for the routed tools: browser, cloud, comms and ml. These tools share a cartridge and are told apart by a routing key derived from the tool name. After this change the tool name alone decides the route, and arguments that a routed tool's inputSchema does not declare are refused with -32602.

📌 New pins

Head SHA: f01e55a. This PR adds or changes no action, lockfile or container pins.

Changes

• mcp-bridge/lib/dispatcher.js:
  • New ROUTED_TOOLS table (tool name → cartridge plus routing key). It replaces four switch arms. The routing key is written after the caller's arguments.
  • New validateRoutedArgs and declaredArgs. The hardening gate refuses arguments that a routed tool's inputSchema does not declare.
  • The gate also refuses non-object arguments.
  • JSDoc added to dispatchTool, hardeningGate and the new helpers.
• Scope: non-routed tools are unchanged. coord_send reads sender_role, which no schema declares, so applying the check to every tool would break it.
• mcp-bridge/tests/routing_args_test.js (new, 16 tests). Each routed tool still reaches its cartridge with the right key. An argument cannot change the routing key. Undeclared arguments are refused. Non-routed handling is unchanged. fetch is stubbed for this file only and restored afterwards, because bun shares one process across files.
• .github/workflows/e2e.yml (node, deno and bun unit lines) and package.json test now include the new file.

RSR Quality Checklist

Required

• Tests pass. node --test: 68/68 across the four bridge test files. bun test: 68/68. deno test: 68/68.
• Code is formatted: no formatter is configured for mcp-bridge/ JS. The code follows the surrounding style by hand.
• Linter is clean: no JS linter runs on mcp-bridge/ locally. CI scanners will report on this PR.
• No banned language patterns. Plain ESM JS, as in the existing bridge. Nothing new in TS, Python or Go. The npm test script line already existed and only gained a file name.
• No unsafe blocks: there is no Rust or Zig in this change.
• No banned functions.
• SPDX headers: the new test file carries MPL-2.0. The modified files keep theirs.
• No secrets, credentials or .env files.

As Applicable

• .machine_readable/* not updated: project state and integrations are unchanged.
• Documentation not updated: the advertised tool schemas are unchanged. Only calls that already violated those schemas are now refused.
• TOPOLOGY.md not updated: architecture is unchanged.
• CHANGELOG / release notes: to follow with the patch release that ships this fix.
• New dependencies: none.
• ABI/FFI: not touched.

Testing

• node --test mcp-bridge/tests/{routing_args,dispatch,http_transport,path_claims}_test.js: 68 pass, 0 fail.
• bun test (same files): 68 pass, 0 fail. deno test --allow-read --allow-env --allow-run --allow-net (same files): 68 passed.
• Control: routing_args_test.js against origin/main's dispatcher.js gives 8 pass and 8 fail. Every routing and undeclared-argument test fails there; the well-formed-call and non-routed tests pass.

🤖 Generated with Claude Code

https://claude.ai/code/session_019j8She9eTFx54r6aL6sCHP

[coderabbitai]

Warning

Review limit reached

You've used all free OSS reviews for now. Wait for the free limit to reset to keep reviewing this public repository.

Next included review available in 40 minutes.

Check out review usage here.

View limit details

Limit details: You’ve used the included review currently available.

Learn how review limits work.

Review configuration:

⚙️ Run configuration

• Configuration used: Organization UI
• Review profile: ASSERTIVE
• Plan: Advanced
• Run ID: 77350813-43f6-4f31-b854-db5b615720fe

📥 Commits

Reviewing files that changed from the base of the PR and between 9180ee6 and f01e55a.

📒 Files selected for processing (4)

• .github/workflows/e2e.yml
• mcp-bridge/lib/dispatcher.js
• mcp-bridge/tests/routing_args_test.js
• package.json

• Autopilot · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

---

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.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[github-actions]

🏁 path-claims bench

Commit 3b8f564

Numbers

path-claims bench  (node v22.23.3)

  scenario                                              iters       ms        ns/op          ops/s
  --------------------------------------------------------------------------------------------------------------
  register: 10 active claims, 3 new paths               50000 iters    181 ms      3.63 µs/op    275.6k ops/s
  register: 100 active claims, 3 new paths              20000 iters    319 ms     15.95 µs/op     62.7k ops/s
  register: 1000 active claims, 3 new paths              5000 iters    973 ms    194.72 µs/op      5.1k ops/s
  register: 100 active claims, 20 new paths              5000 iters    371 ms     74.37 µs/op     13.4k ops/s

  pathsOverlap: deep diverge at segment 4             1000000 iters    155 ms     155.6 ns/op     6.43M ops/s
  pathsOverlap: short prefix match                    1000000 iters    146 ms     146.5 ns/op     6.83M ops/s

  refresh (existing claim)                             100000 iters     11 ms     115.1 ns/op     8.69M ops/s
  list (100 active claims)                              50000 iters    296 ms      5.94 µs/op    168.4k ops/s

  (Bench numbers depend on host; use deltas across commits, not absolute values.)

Host-dependent — compare deltas across commits, not absolute values.

[github-actions]

🔍 Hypatia Security Scan

Findings: 113 issues detected

Severity	Count
🔴 Critical	10
🟠 High	20
🟡 Medium	83

⚠️ Action Required: Critical security issues found!

View findings

[
  {
    "reason": "Job `sonarqube` in build.yml has no `timeout-minutes:` declaration. Default is 6 hours — a stuck codeload fetch or runner hang can burn budget. Add `timeout-minutes: 10` (or proportional).",
    "type": "missing_timeout_minutes",
    "file": ".github/workflows/build.yml",
    "action": "flag",
    "rule_module": "workflow_audit",
    "severity": "medium",
    "recipe_id": "recipe-add-workflow-timeout-minutes",
    "job": "sonarqube"
  },
  {
    "reason": "Job `triage` in label-triage.yml has no `timeout-minutes:` declaration. Default is 6 hours — a stuck codeload fetch or runner hang can burn budget. Add `timeout-minutes: 10` (or proportional).",
    "type": "missing_timeout_minutes",
    "file": ".github/workflows/label-triage.yml",
    "action": "flag",
    "rule_module": "workflow_audit",
    "severity": "medium",
    "recipe_id": "recipe-add-workflow-timeout-minutes",
    "job": "triage"
  },
  {
    "reason": "Job `sync` in labels.yml has no `timeout-minutes:` declaration. Default is 6 hours — a stuck codeload fetch or runner hang can burn budget. Add `timeout-minutes: 10` (or proportional).",
    "type": "missing_timeout_minutes",
    "file": ".github/workflows/labels.yml",
    "action": "flag",
    "rule_module": "workflow_audit",
    "severity": "medium",
    "recipe_id": "recipe-add-workflow-timeout-minutes",
    "job": "sync"
  },
  {
    "reason": "Job `deploy` in pages-deploy.yml has no `timeout-minutes:` declaration. Default is 6 hours — a stuck codeload fetch or runner hang can burn budget. Add `timeout-minutes: 10` (or proportional).",
    "type": "missing_timeout_minutes",
    "file": ".github/workflows/pages-deploy.yml",
    "action": "flag",
    "rule_module": "workflow_audit",
    "severity": "medium",
    "recipe_id": "recipe-add-workflow-timeout-minutes",
    "job": "deploy"
  },
  {
    "reason": "Step uses `peter-evans/repository-dispatch` with `token: ${{ secrets.FARM_DISPATCH_TOKEN }}` but has no `if: secrets.FARM_DISPATCH_TOKEN != ''` gate. On repos where the secret hasn't been propagated the action fails on every push, red-maining the repo. Add the step-level gate (or env+if pattern) so the missing-secret path is a clean skip instead of a red.",
    "type": "secret_action_without_presence_gate",
    "file": ".github/workflows/instant-sync.yml",
    "action": "peter-evans/repository-dispatch",
    "rule_module": "workflow_audit",
    "severity": "high",
    "fix_recipe": "add_secret_presence_gate"
  },
  {
    "reason": "codeql.yml does not list `language: actions` in its matrix, but the repo has workflow files. CodeQL's `actions` language scans workflow YAML for injection and other CI/CD-specific weaknesses — every repo with workflows benefits. Add an entry to `matrix.include` with `language: actions` + `build-mode: none`.",
    "type": "codeql_missing_actions_language",
    "file": ".github/workflows/codeql.yml",
    "action": "flag",
    "rule_module": "workflow_audit",
    "severity": "medium",
    "fix_recipe": "add_codeql_actions_language"
  },
  {
    "line": 39,
    "reason": "job in .github/workflows/labels.yml references `secrets.*` but does not install `step-security/harden-runner` — review outbound-egress monitoring",
    "type": "RE001",
    "file": ".github/workflows/labels.yml",
    "action": "report",
    "rule_module": "research_extensions",
    "severity": "medium"
  },
  {
    "line": 46,
    "reason": "job in .github/workflows/push-email-notify.yml references `secrets.*` but does not install `step-security/harden-runner` — review outbound-egress monitoring",
    "type": "RE001",
    "file": ".github/workflows/push-email-notify.yml",
    "action": "report",
    "rule_module": "research_extensions",
    "severity": "medium"
  },
  {
    "line": 32,
    "reason": "job in .github/workflows/build.yml references `secrets.*` but does not install `step-security/harden-runner` — review outbound-egress monitoring",
    "type": "RE001",
    "file": ".github/workflows/build.yml",
    "action": "report",
    "rule_module": "research_extensions",
    "severity": "medium"
  },
  {
    "line": 44,
    "reason": "job in .github/workflows/container-publish.yml references `secrets.*` but does not install `step-security/harden-runner` — review outbound-egress monitoring",
    "type": "RE001",
    "file": ".github/workflows/container-publish.yml",
    "action": "report",
    "rule_module": "research_extensions",
    "severity": "medium"
  }
]

Powered by Hypatia Neurosymbolic CI/CD Intelligence