Papermark (agent-first CLI)

Use this skill when the task is about Papermark dataroom discovery or analytics.

Default stance:

• Prefer the papermark CLI, not manual browser clicking.
• Prefer --json for machine use.
• Stay read-first unless the CLI explicitly supports a mutation.
• Model dataroom contents through folder traversal, not just top-level documents.
• If papermark is missing from PATH, install the published CLI with npm i -g papermark-cli. If global npm install is unavailable, install from a local checkout with npm link.

Default workflow

• Sanity check: papermark doctor --json
• List rooms: papermark datarooms list --json
• If needed, override the saved workspace with --team-id <id>
• Take the dataroom id from datarooms list and feed it into the follow-up commands
• Inspect one room: papermark datarooms get <id> --json
• Read content tree: papermark datarooms folders <id> --json
• Use --limit for a wider summary and --raw only when the full nested folder tree is truly needed
• Inspect access model: papermark datarooms links <id> --json, papermark datarooms groups <id> --json, papermark datarooms permission-groups <id> --json
• Inspect analytics: papermark datarooms views <id> --json
• Inspect analytics summary counts: papermark datarooms views-count <id> --json
• Inspect viewers: papermark datarooms viewers <id> --json
• Inspect room stats: papermark datarooms stats <id> --json
• Inspect export jobs: papermark datarooms export-visits <id> --json

Worked example

If the user asks "show me the analytics and folder structure for the investor dataroom":

1. Run papermark datarooms list --json
2. Pick the matching dataroom id from data.datarooms
3. Run papermark datarooms folders <id> --json for the content tree
4. Run papermark datarooms views <id> --json, papermark datarooms viewers <id> --json, or papermark datarooms stats <id> --json depending on the level of analytics detail needed

Auth

If auth is missing:

• Best interactive path: papermark auth login
• If you already have a visible Chrome session with remote debugging: papermark auth login --cdp-port 9333
• Saved local config: printf '%s' '{"sessionToken":"...","currentTeamId":"..."}' | papermark auth set --stdin
• Best ephemeral path: set PAPERMARK_SESSION_TOKEN and PAPERMARK_CURRENT_TEAM_ID
• If the stored team id is wrong, pass --team-id <id> on the dataroom command instead of recapturing auth

Avoid pasting full session tokens into logs or chat. Treat dataroom, analytics, links, groups, and viewer output as sensitive workspace data.

Important constraints

• This adapter targets a private authenticated surface, not a documented public management API.
• Dataroom content is folder-first.
• export-visits currently inspects export jobs and does not yet start them.

Contract essentials

• Prefer --json for agent work.
• With --json, stdout should contain exactly one JSON object.
• Progress and status belong on stderr.
• Exit codes:
  • 0 success
  • 1 request failure, upstream failure, or failed checks
  • 2 auth/user action required or invalid input
• Common error codes:
  • AUTH_MISSING
  • AUTH_INVALID
  • NOT_FOUND
  • RATE_LIMITED
  • UPSTREAM_5XX
  • TIMEOUT
  • VALIDATION
  • CHECK_FAILED