Dungbeetle MCP server

Read from the package source without running it. The installed server may list more.

• dungbeetle_approve_runwrite action

  Approve a run. With promote=true, the run's candidate snapshots become new hosted baseline versions — this MUTATES baselines and should be confirmed by a human first. Returns the recorded review and any promoted targets.

• dungbeetle_get_analytics

  Pass rate, run trend, and per-target flakiness over the recent run window. Use to spot flaky targets before approving.

• dungbeetle_get_baseline_history

  Full version history for one baseline target (versions, digests, timestamps).

• dungbeetle_get_repository

  Return the repository the configured credentials resolve to (id, name). Everything else is scoped to it.

• dungbeetle_get_review_url

  Deep link to the web review page for a run, where a human reviews diffs and approves baselines.

• dungbeetle_get_runwrite action

  Structured summary of a run for triage: per-target status, counts, changed targets, current review state, and a review URL. Use detail="full" to include the text diff per target. Screenshots are never inlined — open the review URL for images.

• dungbeetle_get_usage

  Current-period usage (snapshots ingested, storage bytes, repositories) alongside the plan limits it is measured against, plus any active quota warnings. Use to check quota before pushing runs or baselines.

• dungbeetle_health

  Liveness/version probe for the configured Dungbeetle server. Use to confirm connectivity before other calls.

• dungbeetle_list_baselines

  List the repository's baseline targets with their latest version and last-updated time. Start here to see what's under test.

• dungbeetle_list_runs

  List recent runs for the repository, newest first — use this to discover a run id to triage with dungbeetle_get_run. Returns compact per-run summaries (id, status/passed, mode, branch, commit, createdAt, reviewState, failedCount, url); no diffs, no images.

• dungbeetle_reject_runwrite action

  Reject a run's changes (records a rejected review in the append-only audit trail). Does not modify baselines.

• dungbeetle_reopen_runwrite action

  Reopen a previously approved or rejected run by setting its review state back to pending, so it can be re-reviewed. Recorded in the append-only audit trail; does not modify baselines.

• Code scan7 source files scanned25/25
• –Live reliabilityno gateway calls yet and no remote to proben/a
• –Tool poisoningtools not inspected (local package is not executed); not countedn/a
• Auth qualitystatic API keys via environment variables6/15
• Maintenancerepository not readable: repo not found3/15
• Maintainer identityverified namespace with website, no repo4/10

Safety score

58/100

higher than 13% of 48,255 servers

Tools

12

the typical server has 7

Tools that change things

33%

typical: 0%. Fewer write tools means less an agent can break

Downloads a week

20

more than 7% of 12,191 servers with a package

Dungbeetle: common questions

Is Dungbeetle MCP server safe?

With care: it is graded C, so read the findings first (58/100). Read the Dungbeetle safety report

How do I install Dungbeetle?

It runs on your machine. Copy the Claude Code, Cursor, VS Code or Claude Desktop config from the install section.

Does Dungbeetle need an API key?

Yes. The registry entry asks for DUNGBEETLE_AGENT_TOKEN, DUNGBEETLE_CLIENT_SECRET.

Is Dungbeetle maintained?

The latest release is v0.1.0.

What can I use instead of Dungbeetle?

Servers from other publishers that do the same job: Sem MCP server.

How does Dungbeetle compare with other MCP servers?

safety score 58/100 (higher than 13% of 48,255 servers); tools 12 (the typical server has 7); tools that change things 33% (typical: 0%. Fewer write tools means less an agent can break). Full comparison.