Elnora Platform

Elnora Platform CLI

Route Elnora Platform queries to the correct sub-skill. Load only what is needed.

Tool Access

Elnora is available via two methods. Use whichever is configured.

Option A — CLI via Bash (preferred)

Run commands via your Bash/Shell tool as elnora <group> <action> .... Verify with elnora --version. CLI uses ~5× fewer tokens than MCP.

Option B — MCP tools (when CLI isn't installed)

Look for tools prefixed mcp__elnora__ in your available tools. Each domain-specific sub-skill (elnora-projects, elnora-tasks, etc.) has an MCP Tool Names table showing the mapping from CLI commands to MCP tool names.

If neither is available, tell the user to install one:

• CLI: curl -fsSL https://cli.elnora.ai/install.sh | bash (macOS/Linux) or irm https://cli.elnora.ai/install.ps1 | iex (Windows)
• MCP: claude mcp add elnora --transport http --scope user https://mcp.elnora.ai/mcp then /mcp to authenticate.

Never fabricate tool names. This skill routes to sub-skills; each sub-skill lists its valid commands and their MCP equivalents.

Troubleshooting

If anything looks wrong — auth errors, unexpected hosts, or connectivity failures — run elnora doctor first. It reports the actual API / AI-server / MCP hosts, auth state, version, and connectivity. Run elnora config show to print the resolved endpoints and active profile. Do not guess hostnames or URLs.

Invocation

CLI="elnora"

Global flags go BEFORE the subcommand (recommended, always works):

$CLI --compact projects list            # recommended
$CLI projects list --compact            # also works for most commands

Global Flags

Updates

On first use in a session, run elnora update to check for updates. If an update is available, inform the user and follow the printed instructions to update.

Auth

Requires ELNORA_API_KEY environment variable (prefix: elnora_live_), ELNORA_MCP_API_KEY, or a saved profile in ~/.elnora/profiles.toml.

$CLI --compact auth status
# -> {"profile":"default","authenticated":true,"projectCount":N}

Get keys: platform.elnora.ai > Settings > API Keys

Routing Table

Positional Argument Convention

Required string fields ending in Id (e.g. taskId, projectId, fileId) become positional arguments. Everything else becomes a flag.

elnora tasks get <task-id>              # taskId -> positional
elnora tasks list --project <UUID>      # project -> flag (doesn't end in "Id")
elnora files fork <file-id> --target-project <UUID>  # fileId -> positional, targetProject -> flag

Rule: positional if and only if the field is: required + not boolean + no enum choices + key ends in "Id".

ID Format

All IDs are UUIDs: bfdc6fbd-40ed-4042-9ea7-c79a5ec90085. Invalid format exits 1 with a validation error to stderr.

Exception: account get and account update use userId which accepts any string (typically an integer like 42).

Pagination

List endpoints return:

{"items":[...],"page":1,"pageSize":25,"totalCount":N,"totalPages":N,"hasNextPage":true,"hasPreviousPage":false}

Use --page N --page-size N (max 100). Check hasNextPage to paginate.

Message endpoints use cursor-based pagination: check hasMore and pass --cursor <nextCursor>.

Error Contract

Errors -> stderr, exit code > 0:

{"error":"message","code":"AUTH_FAILED","suggestion":"how to fix"}

Rate Limits

HTTP 429 on limit. Check the Retry-After header for seconds to wait.

Agent strategy: On exit code 5, read Retry-After from stderr if available, otherwise wait 60 seconds before retrying. Do not retry in a tight loop.

File Upload Limits

Common Workflow

Projects contain tasks and files. Typical flow:

1. --compact --fields "id,name" projects list -> list projects, pick one by name
2. tasks create --project <ID> --message "..." --stream -> create task and stream the agent response
3. tasks send <TASK_ID> --message "..." --stream -> send follow-up and stream response
4. tasks messages <TASK_ID> -> read conversation history (or use --wait instead of --stream in steps 2-3)
5. files list --project <ID> -> browse generated outputs
6. files content <FILE_ID> -> read a protocol file

Project ID: List projects once (step 1), pick the one matching the user's context, and reuse the ID for all subsequent commands. Don't re-list projects for every task.

See elnora-tasks skill for full response retrieval details (--stream, --wait, MCP behavior).

All Command Groups

elnora account      Manage user account, agreements, and legal docs
elnora api-keys     Manage API keys and creation policy
elnora audit        View audit logs
elnora auth         Manage authentication
elnora completion   Generate shell completion script
elnora doctor       Run diagnostics (API, auth, version, config checks)
elnora feedback     Submit feedback
elnora files        Manage files (share, move, versions, batch upload)
elnora flags        Manage global feature flags (SystemAdmin)
elnora folders      Manage Knowledge Base folders (browse, share)
elnora health       Check platform reachability
elnora library      Manage organization library
elnora mcp          Run as MCP server (--http or --stdio)
elnora open         Open platform pages in browser
elnora orgs         Manage organizations (members, directory, set-autotidy, billing)
elnora projects     Manage projects
elnora review       Approve/reject Knowledge Base review-queue items
elnora search       Search tasks, files, and file content
elnora setup        Configure AI coding tools (Claude Code, Cursor, VS Code, Codex)
elnora tasks        Manage tasks
elnora update       Check for updates (use --install to apply)
elnora whoami       Show current profile and org