Command reference: OpenPort

Overview

OpenPort commands let an automation read workspace data and propose or execute actions within the permissions of its Agent Key. Data commands need an Agent Key; review commands need a person’s browser sign-in. All of them need a server that supports OpenPort, and review commands also need approvals support.

Read data (Agent Key)

figena openport manifest

Shows the tool manifest: the actions and data available to your Agent Key.

figena openport workspaces

Lists the workspaces your Agent Key can reach.

figena openport transactions

Lists transactions in a workspace. The workspace comes from --workspace <id> (or --org <id>) or the profile’s orgId.

• --kind <kind> and --search <text>: filter the list.
• --from <date> and --to <date>: limit the date range.
• --sort <order>, --page <n>, and --page-size <n>: order and page through results.

Propose and execute actions (Agent Key)

figena openport preflight

Checks an action before you submit it and returns its expected impact with an impactHash. --action <name> is required; pass the payload with --payload <file> or --payload-json '<json>'.

figena openport action

Creates an action from --action <name> and a payload. Every action is recorded as a draft first. Add --execute to ask Figena to run it right away, together with --justification "<reason>", which is saved with the action. Whether it runs depends on the auto-execute settings your workspace has for the Agent Key; when execution isn’t allowed, the action stays a draft for review and the result explains why.

• --preflight-hash <hash>: the impactHash from openport preflight. High-risk actions usually need it to execute, and it must still match the impact when you submit.
• --force-draft: submit the action as a draft for review.
• --request-id <id> and --idempotency-key <key>: identify the request. With an idempotency key, a request that doesn’t execute may be retried once after a temporary failure; executed actions are never retried by the CLI.

figena openport preflight --action <action-name> --payload ./payload.jsonfigena openport action --action <action-name> --payload ./payload.json --force-draft

figena openport draft get <draftId>

Shows one draft.

Review drafts and export audits (browser sign-in)

figena openport drafts list

Lists drafts awaiting review. Filter with --app <appId>, --org <id>, and --status <status>, and page with --page-size <n> (or --limit) and --cursor <cursor>. Add --all to follow every page, and --max-items <n> to stop after that many drafts. The result includes items, nextCursor, and the number of pages read.

figena openport draft approve <draftId>

Approves a draft. Add --note "<text>" to record why.

figena openport draft reject <draftId>

Rejects a draft. Add --note "<text>" to record why.

figena openport audit export

Exports the OpenPort audit log as CSV. Filter with --app, --org, --from, --to, and --limit. With --file <path>, the CLI writes the CSV to that file and prints “Exported audit CSV to” followed by the path, together with the number of rows; without it, the CSV is returned in the result data.

figena openport audit export --org <workspace-id> --from <date> --to <date> --file ./openport-audit.csv