Skip to main content
Glama

Ship Signal

Is this release ready to ship? Ship Signal answers that for any Jira fix version, using the context your team already has in Jira and Confluence, and it shows its working.

It is a Forge app with two surfaces:

  1. A Jira project page (UI Kit). It lists the project's unreleased versions, scores the selected one from 0 to 100, gives a verdict (Ready, At risk or Not ready), and shows a table of every check with the points it cost and the issues behind it, each linked.

  2. Rovo MCP tools (rovo:mcp module). The same checks are exposed as tools, so a Rovo agent (or any MCP-compatible client the site admin connects) can answer "Is 2.4 ready to ship?" or "What's blocking 2.4?".

How readiness is scored

The score starts at 100. Each check removes points, and each one has a cap so a single problem can't hide the others. Every deduction is shown with its reason.

Check

Deduction

Open high-priority issues (Highest, High, Blocker, Critical)

-10 each, max -30

Flagged issues, issues blocking open work, or issues blocked by unresolved issues

-8 each, max -25

Open issues with no assignee

-3 each, max -15

Issues reopened in the last 14 days (resolution cleared)

-5 each, max -10

No Confluence release-notes page linked to the version

-10

Schedule risk: past the release date with work open, or ≤7 days left with >25% open

-10

Verdict: Ready at 85+ with no open high-priority issues, At risk at 60+, otherwise Not ready.

The release-notes check looks for a Confluence (/wiki/) link in the version's remote links, then in the version description.

Related MCP server: jev-agent-quality

Rovo MCP tools

Tool

Inputs

Returns

get-release-readiness

projectKey, version (name or id)

Score, verdict and every check as text

list-release-blockers

projectKey, version

Each blocking issue with assignee and reason

list-upcoming-releases

projectKey

Each unreleased version with verdict, score and open count

Example prompts for a Rovo agent that has these tools:

  • "Is SHIP version 2.4.0 ready to ship?"

  • "What's blocking the 2.4.0 release in SHIP, and who owns each blocker?"

  • "Which upcoming SHIP releases are at risk?"

All Jira calls run as the invoking user (asUser()), so the page and the tools only see what that person can already see. The only scope is read:jira-work. The app never writes to Jira.

Install

npm install
npm test            # scoring unit tests
npm run typecheck
forge deploy -e development
forge install --site <your-site>.atlassian.net --product jira -e development

Then open a Jira project and choose Ship Signal in the project sidebar. To use the tools, create a custom agent in Rovo Studio, add the Ship Signal tools from Connected apps, and publish it.

Code

  • src/core/readiness.ts: pure scoring logic, unit-tested in src/core/readiness.test.ts

  • src/core/jira.ts: gathers versions, issues, links, flags, reopen history and release-notes links

  • src/resolvers.ts: backend for the project page

  • src/mcp.ts: the Rovo MCP tool handlers

  • src/frontend/index.tsx: the UI Kit project page

License

MIT. See LICENSE.

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    Related MCP Connectors

    Related MCP Servers

    • A
      license
      Not graded
      quality
      D
      maintenance
      MCP server for auditing e-commerce stores' AI agent readiness, providing deterministic scores, blockers, and fixes. Enables agents to call check_merchant_readiness, get_readiness_report, and list_agent_ready_stores.
      MIT
    • A
      license
      Not graded
      quality
      C
      maintenance
      Exposes a spec-compliant JSON-RPC 2.0 endpoint that lets external AI IDEs and agents query live Jira and Confluence context, traverse multi-hop dependency and policy graphs to simulate a release's blast radius, and mint tamper-evident SHA-256 Merkle release receipts. It also supports triggering autonomous remediation actions that create mitigation subtasks, alert affected downstream teams, and raise a ship-confidence score before deployment.
      11 npm
      MIT