Forge Connector

Forge Connector

Builds a graph:connector Forge app that ingests external data into Atlassian's Teamwork Graph so it appears in Rovo Search and Rovo Chat.

Critical Rules

1. Must install in Jira — Apps using Teamwork Graph modules must be installed on a Jira site. Confluence-only installs will not work.
2. Never ask for credentials in chat — Direct users to run forge login in their own terminal.
3. Always run the scaffold script yourself — Do not only give manual instructions; run scripts/scaffold_connector.py to generate the boilerplate.
4. Always ask the user for their Atlassian site URL when install is needed — never discover or guess it.
5. Atlassian deletes data on disconnect — When action = 'DELETED', the app only needs to clean up local state; Atlassian removes the Teamwork Graph data automatically.
6. Handler arguments are passed directly — Forge passes the request object as the first argument to handlers, NOT nested under event.payload. Config values are at request.configProperties, NOT event.payload.config. This is the most common source of TypeError: Cannot destructure property of undefined errors.
7. Use @forge/kvs for storage — Import kvs from @forge/kvs. Do NOT use @forge/storage — its storage export is undefined at runtime in connector functions.
8. Use graph named export from @forge/teamwork-graph — The correct import is const { graph } = require('@forge/teamwork-graph'). Call graph.setObjects({ objects, connectionId }). Do NOT import setObjects as a named export directly.
9. validateConnectionHandler must return { success, message } — Do NOT throw an Error. Return { success: false, message: '...' } to reject, { success: true } to accept.
10. function declarations belong under modules — In manifest.yml, function: is a key under modules:, not a top-level key. Placing it at the top level causes a lint error.
11. formConfiguration uses form array with type: header — Do NOT use fields: or beforeYouBegin:. The correct format uses form: [{ key, type: header, title, description, properties: [...] }].
12. Scopes are read/write/delete:object:jira — Use read:object:jira, write:object:jira, delete:object:jira. The scopes read:graph:teamwork and write:graph:teamwork are invalid and will fail forge lint.

MCP Prerequisites

Agent Workflow — Complete Steps 0–5 in Order

Step 0: Prerequisites

Check Node.js (node -v, requires 22+), Forge CLI (forge --version), and login (forge whoami). Install missing tools:

npm install -g @forge/cli

Tell the user to run forge login in their terminal if not authenticated.

Step 1: Discover Developer Spaces

Note: forge developer-spaces list does NOT exist in Forge CLI 12.x. You cannot list developer spaces non-interactively.

forge create requires an interactive TTY to select a developer space. Ask the user to run it themselves:

Tell the user:
  cd <parent-directory>
  forge create --template blank <app-name>

  When prompted, select a Developer Space and let it complete.
  Come back when done.

The --dev-space-id flag in the scaffold script is optional and can be omitted — the script has been updated to skip it when not provided.

Step 2: Scaffold the Connector App

Run from the skill directory (the directory containing this SKILL.md). --dev-space-id is optional:

python3 -m scripts.scaffold_connector \
  --name <app-name> \
  --connector-name "<Human Readable Name>" \
  --object-type atlassian:document \
  --directory <parent-directory>

Add --dev-space-id <id> only if you have the ID from a previous step.

Object type selection — pick the type that best matches the content being ingested (see Object Types table). For mixed content, use atlassian:document as the default.

Form config flag — add --has-form-config if the admin must provide API credentials or connection details (typical for external systems). Omit it for apps that operate entirely within Atlassian (no external credentials needed).

If scaffold fails because forge create needs a TTY: The scaffold script will print a manual fallback command. Have the user run forge create interactively, then continue from Step 3 — the scaffold script only needs to write manifest.yml and src/index.js after the directory exists.

Step 3: Customize the Generated Code

After scaffolding (or after the user runs forge create interactively):

cd <app-name>
npm install

The blank template generates src/index.js (JavaScript, not TypeScript). Edit it to add your API calls. The scaffold generates working handler skeletons; fill in your business logic.

Key files to edit

setObjects — ingest data into Teamwork Graph

Use the graph named export — do NOT destructure setObjects directly:

const { graph } = require('@forge/teamwork-graph');

const result = await graph.setObjects({
  connectionId,          // required — the connectionId from the handler request
  objects: [
    {
      schemaVersion: '1.0',
      id: 'unique-id-from-source',      // unique per connectionId
      updateSequenceNumber: 1,
      displayName: 'My Document Title',
      url: 'https://source-system.example.com/doc/123',
      createdAt: '2024-01-15T10:00:00Z',        // ISO 8601
      lastUpdatedAt: '2024-01-20T14:30:00Z',
      permissions: [{
        accessControls: [{
          principals: [{ type: 'EVERYONE' }],   // or restrict to specific users
        }],
      }],
      'atlassian:document': {
        type: {
          category: 'DOCUMENT',   // see Document Categories table below
          mimeType: 'application/vnd.google-apps.document',
        },
        content: {
          mimeType: 'application/vnd.google-apps.document',
          text: 'document title or snippet for search indexing',
        },
      },
    },
  ],
});

if (!result.success) {
  console.error('setObjects error:', result.error);
}

• Max 100 objects per call — batch large datasets with a loop
• id must be unique per connectionId
• connectionId is required in every graph.setObjects() call

Document Categories (for atlassian:document.type.category)

getObjectByExternalId — look up a single object

const { graph } = require('@forge/teamwork-graph');

const data = await graph.getObjectByExternalId({
  externalId: 'unique-id-from-source',
  objectType: 'atlassian:document',
  connectionId,
});
if (data.success) console.log(data.object);

Step 4: Deploy and Install

You MUST run the deploy script — do not only give the user manual forge deploy commands.

The deploy script lives in the forge-app-builder skill, not in this skill. Derive its directory from the path of this SKILL.md: go up two levels (skills/forge-connector/ → skills/) then into forge-app-builder/. Run all commands below from that directory.

# Derive forge-app-builder skill dir from this SKILL.md's path:
# e.g. if this file is at /path/to/skills/forge-connector/SKILL.md
# then the deploy script dir is: /path/to/skills/forge-app-builder/

# If you have the site URL:
python3 -m scripts.deploy_forge_app \
  --app-dir <app-directory> \
  --site <site-url> \
  --product jira

# If you don't have the site URL yet, deploy first then ask:
python3 -m scripts.deploy_forge_app \
  --app-dir <app-directory> \
  --product jira \
  --deploy-only
# Ask: "What is your Atlassian site URL (e.g. yourcompany.atlassian.net)?"
python3 -m scripts.deploy_forge_app \
  --app-dir <app-directory> \
  --site <site-url> \
  --product jira \
  --skip-deps

Step 5: Connect via Atlassian Administration

After deployment, tell the user to:

1. Go to Atlassian Administration → Apps → [site] → Connected apps
2. Find the app → View app details → Connections tab
3. Click Connect under the connector
4. Fill in any configuration fields (if formConfiguration was defined)
5. Click Connect — this triggers onConnectionChange with action: CREATED and starts data ingestion

Step 6: Monitor with forge tunnel

Use forge tunnel during development to stream live logs directly to your terminal as the connector functions execute. This is the fastest way to catch errors in onConnectionChangeHandler, validateConnectionHandler, and setObjects calls without waiting for forge logs.

Tell the user to run this in their own terminal (it requires an interactive session):

cd <app-directory>
forge tunnel

With the tunnel active, any invocation of the connector functions (e.g. clicking "Connect" in Atlassian Admin, or triggering a scheduled re-ingestion) will stream output immediately. Look for:

• [connector] Fetched N items — confirms fetchExternalData() ran
• [connector] Batch 1: N accepted, 0 rejected — confirms setObjects succeeded
• Any uncaught errors or thrown exceptions from validateConnectionHandler

If the tunnel is not running, use forge logs instead to inspect past invocations:

# Most recent 50 log lines from development environment
forge logs -e development --limit 50

# Production logs for a specific site
forge logs -e production --site <your-site> --limit 50

Tunnel vs logs — when to use which:

Note: forge tunnel must be run by the user in an interactive terminal — do not attempt to run it via the agent.

Manifest Reference

Key rules:

• Scopes are read:object:jira, write:object:jira, delete:object:jira — NOT read:graph:teamwork / write:graph:teamwork (those fail forge lint)
• function: is declared under modules:, not at the top level
• Egress uses address: not a bare string (run forge lint --fix to auto-correct)
• formConfiguration uses form: [{ type: header, properties: [...] }] — NOT fields: or beforeYouBegin:

Minimal connector (no admin config, no OAuth)

Use when the app operates entirely within Atlassian — no external credentials needed.

app:
  id: <generated-by-forge-create>
  runtime:
    name: nodejs24.x
    memoryMB: 256
    architecture: arm64

permissions:
  scopes:
    - read:object:jira
    - write:object:jira
    - delete:object:jira
    - storage:app

modules:
  graph:connector:
    - key: my-connector
      name: My Service
      icons:
        light: https://cdn.example.com/logo.png
        dark: https://cdn.example.com/logo.png
      objectTypes:
        - atlassian:document
      datasource:
        onConnectionChange:
          function: on-connection-change

  function:
    - key: on-connection-change
      handler: index.onConnectionChangeHandler

Connector with admin form config (API key / URL)

Use when the admin must provide credentials to connect to an external system.

app:
  id: <generated-by-forge-create>
  runtime:
    name: nodejs24.x
    memoryMB: 256
    architecture: arm64

permissions:
  scopes:
    - read:object:jira
    - write:object:jira
    - delete:object:jira
    - storage:app
  external:
    fetch:
      backend:
        - address: 'https://api.your-service.com'   # note: address: not a bare string

modules:
  graph:connector:
    - key: my-connector
      name: My Service
      icons:
        light: https://cdn.example.com/logo.png
        dark: https://cdn.example.com/logo.png
      objectTypes:
        - atlassian:document
      datasource:
        formConfiguration:
          form:                          # use form:, NOT fields: or beforeYouBegin:
            - key: connectionDetails
              type: header
              title: Connection Details
              description: >
                Provide your My Service API credentials.
                Find them in My Service → Settings → API.
              properties:
                - key: apiKey           # camelCase keys — accessed as request.configProperties.apiKey
                  label: API Key
                  type: string
                  isRequired: true
                - key: apiUrl
                  label: API URL
                  type: string
                  isRequired: true
          validateConnection:
            function: validate-connection
        onConnectionChange:
          function: on-connection-change

  function:                              # function: is under modules:, NOT top-level
    - key: on-connection-change
      handler: index.onConnectionChangeHandler
    - key: validate-connection
      handler: index.validateConnectionHandler

Handler Signatures

Critical: Forge passes the request directly as the first argument — it is NOT wrapped under event.payload. Config form values are at request.configProperties, not event.payload.config. Getting this wrong causes TypeError: Cannot destructure property of undefined.

onConnectionChange

const { kvs } = require('@forge/kvs');
const { graph } = require('@forge/teamwork-graph');

exports.onConnectionChangeHandler = async (request) => {
  // request.action, request.connectionId, request.configProperties
  const { action, connectionId, configProperties } = request;

  if (action === 'DELETED') {
    // Atlassian removes Teamwork Graph data automatically on disconnect.
    // Only clean up locally stored credentials.
    await kvs.deleteSecret(connectionId);
    return { success: true };
  }

  // CREATED or UPDATED — persist credentials and ingest data
  await kvs.setSecret(connectionId, configProperties);
  await ingestAllData(connectionId, configProperties);
  return { success: true };
};

validateConnection

const { fetch } = require('@forge/api');

exports.validateConnectionHandler = async (request) => {
  // request.configProperties — NOT event.payload.config
  const { configProperties } = request;

  // Return { success: false, message } to reject — do NOT throw an Error.
  // Return { success: true } to accept.
  const response = await fetch(`${configProperties['apiUrl']}/health`);
  if (!response.ok) {
    return { success: false, message: 'Invalid API credentials. Please check your settings.' };
  }
  return { success: true, message: 'Connection validated successfully.' };
};

refreshIngestion (scheduled trigger)

exports.refreshIngestionHandler = async () => {
  const activeConnections = await kvs.get('active-connections') ?? [];
  for (const connectionId of activeConnections) {
    const config = await kvs.getSecret(connectionId);
    if (config) await ingestAllData(connectionId, config);
  }
};

Object Types

Objects in bold are indexed in Rovo Search and Rovo Chat.

Rovo Search / Rovo Chat Surfacing

Once ingested:

• Objects appear in Rovo Search under a subfilter named after the connector's nickname (set by admin at connection time)
• Rovo Chat can reference and cite connector objects in responses when queried about topics related to the ingested content
• Data is not available immediately — allow a few minutes for indexing after onConnectionChange fires

To verify ingestion is working:

1. Open Rovo Search on the Jira site
2. Search for text that appears in an ingested object's name or properties
3. Filter by the connector nickname to narrow results

Batching Pattern for Large Datasets

const { graph } = require('@forge/teamwork-graph');

const BATCH_SIZE = 100;

async function ingestAllData(connectionId, config) {
  const items = await fetchExternalData(config);

  for (let i = 0; i < items.length; i += BATCH_SIZE) {
    const batch = items.slice(i, i + BATCH_SIZE);
    const result = await graph.setObjects({
      connectionId,          // required in every call
      objects: batch.map(item => ({
        schemaVersion: '1.0',
        id: item.id,                      // unique per connectionId
        updateSequenceNumber: 1,
        displayName: item.title,
        url: item.url,
        createdAt: item.createdAt,
        lastUpdatedAt: item.updatedAt,
        permissions: [{
          accessControls: [{ principals: [{ type: 'EVERYONE' }] }],
        }],
        'atlassian:document': {
          type: { category: 'DOCUMENT', mimeType: item.mimeType },
          content: { mimeType: item.mimeType, text: item.title },
        },
      })),
    });
    if (!result.success) {
      console.error(`[connector] setObjects error in batch ${Math.floor(i / BATCH_SIZE) + 1}:`, result.error);
    }
  }
}

Scheduled Re-Ingestion (optional)

To keep data fresh, add a scheduled trigger that re-runs ingestion periodically:

# In manifest.yml — under modules:
scheduledTrigger:
  - key: refresh-trigger
    function: refresh-ingestion
    interval: day   # prefer 'day' or 'hour'; avoid 'fiveMinutes'

# Under function:
  - key: refresh-ingestion
    handler: index.refreshIngestionHandler

const { kvs } = require('@forge/kvs');

// Track active connections in onConnectionChangeHandler:
//   await kvs.set('active-connections', [...activeConnections, connectionId]);
//   await kvs.setSecret(connectionId, configProperties);  // store credentials securely

exports.refreshIngestionHandler = async () => {
  const activeConnections = await kvs.get('active-connections') ?? [];
  for (const connectionId of activeConnections) {
    const config = await kvs.getSecret(connectionId);  // retrieve stored credentials
    if (config) await ingestAllData(connectionId, config);
  }
};

Scripts

The scaffold script is in this skill's directory. The deploy script is in the forge-app-builder skill directory — always cd there (or derive the path from this SKILL.md's location) before running it.

Troubleshooting

Naming and Logo Guidelines

• Use the official service name as the connector name (e.g. Google Drive, not Drive Connector by Acme)
• Use the official service logo for icons — do not modify or combine with your own branding
• These guidelines apply only to the graph:connector module; your Forge app itself may use your own branding