@triagly/cli

Query Triagly feedback data from the command line or AI agents.

Prerequisites

• Node.js >= 18

Installation

npm install -g @triagly/cli
# or
npx @triagly/cli <command>

Quick Start

# 1. Authenticate (opens browser)
triagly login

# 2. List recent feedback
triagly list

# 3. Filter and pretty-print
triagly list --status new --classification bug --pretty

Authentication

Interactive login (recommended)

triagly login

Opens your browser for GitHub OAuth. After authenticating, the CLI creates a tsk_live_... API key for your project and saves it to ~/.triagly/config.json. You won't need to do this again unless you log out.

triagly logout   # remove stored credentials

Environment variable (CI / headless)

export TRIAGLY_API_KEY="tsk_live_..."
export TRIAGLY_PROJECT_ID="<uuid>"

triagly list

Per-command flag

triagly list --api-key tsk_live_... --project <uuid>

Priority order (highest wins): --api-key flag → TRIAGLY_API_KEY env → ~/.triagly/config.json

Switching Projects

If your account has multiple projects, triagly login selects the first one. To switch:

triagly use <project-id>

Or set TRIAGLY_PROJECT_ID in your environment.

Commands

triagly list

List feedback items for the active project.

triagly list [options]

| Option | Description | Default | |--------|-------------|---------| | --status <value> | Filter by status: new, in_progress, resolved, dismissed | — | | --classification <value> | Filter by AI classification: bug, feature, improvement, question | — | | --sentiment <value> | Filter by AI sentiment: critical, high, medium, low | — | | --source-channel <value> | Filter by source channel | — | | --tags <csv> | Filter by tags (comma-separated) | — | | --since <date> | ISO date lower bound (e.g. 2026-01-01) | — | | --until <date> | ISO date upper bound | — | | --limit <n> | Results per page (1–100) | 20 | | --cursor <token> | Pagination cursor from previous response | — | | --sort <field> | created_at or ai_priority_score | created_at | | --order <dir> | asc or desc | desc | | --project <id> | Override active project | — | | --api-key <key> | Override stored API key | — | | --pretty | Human-readable table output | — |

Examples:

# Critical bugs, newest first
triagly list --classification bug --sentiment critical --pretty

# Open items since January
triagly list --status new --since 2026-01-01

# Paginate
triagly list --limit 10
triagly list --limit 10 --cursor <token-from-previous-response>

# Sorted by AI priority score
triagly list --sort ai_priority_score --order desc --limit 5

Output (JSON by default):

{
  "data": [
    {
      "id": "...",
      "title": "Login button broken on Safari",
      "status": "new",
      "ai_classification": "bug",
      "ai_sentiment": "critical",
      "ai_priority_score": 0.92,
      "tags": ["auth", "safari"],
      "source_channel": "widget",
      "created_at": "2026-02-20T10:00:00Z"
    }
  ],
  "meta": {
    "total": 142,
    "limit": 20,
    "has_more": true,
    "next_cursor": "MjAyNi0wMi0yMFQxMDowMDowMFo6YWJjZA=="
  }
}

triagly get <feedback-id>

Fetch a single feedback item with full detail.

triagly get <uuid> [--project <id>] [--api-key <key>] [--pretty]

Returns id, title, description, status, ai_* fields, tags, reporter_email, source_channel, metadata, screenshot_url, timestamps.

triagly stats

Aggregate statistics for the active project.

triagly stats [--since <date>] [--until <date>] [--project <id>] [--api-key <key>] [--pretty]

Output:

{
  "period": { "since": "2026-01-01", "until": "2026-02-24" },
  "total": 312,
  "by_status": { "new": 142, "in_progress": 58, "resolved": 100, "dismissed": 12 },
  "by_classification": { "bug": 89, "feature": 110, "improvement": 88, "question": 25 },
  "by_sentiment": { "critical": 12, "high": 45, "medium": 180, "low": 75 },
  "by_source_channel": { "widget": 200, "email": 80, "csv": 32 },
  "avg_priority_score": 0.61,
  "top_tags": [["auth", 34], ["performance", 28], ["mobile", 22]]
}

triagly export

Export feedback to JSON or CSV (max 5,000 rows).

triagly export [options]

| Option | Description | Default | |--------|-------------|---------| | --format <fmt> | json or csv | json | | --output <file> | Write to file instead of stdout | stdout | | --status <value> | Same filters as list | — | | --classification <value> | | — | | --sentiment <value> | | — | | --since <date> | | — | | --until <date> | | — | | --project <id> | | — | | --api-key <key> | | — |

Examples:

# CSV to file
triagly export --format csv --output ./feedback.csv

# JSON filtered export
triagly export --format json --status new --since 2026-01-01 --output ./new-feedback.json

# Pipe to jq
triagly export --format json | jq '[.[] | select(.ai_priority_score > 0.8)]'

AI Agent Usage

The CLI is designed to be called by AI agents as a subprocess. All output goes to stdout as JSON; errors go to stderr as JSON. This makes it easy to parse in any language.

Claude (tool use / subprocess)

import subprocess, json

result = subprocess.run(
    ["triagly", "list", "--status", "new", "--sentiment", "critical", "--limit", "10"],
    capture_output=True, text=True
)
data = json.loads(result.stdout)
print(f"{data['meta']['total']} critical items")

OpenAI function calling

Define the tool:

{
  "name": "list_feedback",
  "description": "List feedback from Triagly",
  "parameters": {
    "type": "object",
    "properties": {
      "status": { "type": "string", "enum": ["new", "in_progress", "resolved", "dismissed"] },
      "classification": { "type": "string", "enum": ["bug", "feature", "improvement", "question"] },
      "limit": { "type": "integer", "minimum": 1, "maximum": 100 }
    }
  }
}

Execute:

args = ["triagly", "list"]
if params.get("status"): args += ["--status", params["status"]]
if params.get("classification"): args += ["--classification", params["classification"]]
if params.get("limit"): args += ["--limit", str(params["limit"])]

result = subprocess.run(args, capture_output=True, text=True)

CI / GitHub Actions

- name: Export feedback report
  env:
    TRIAGLY_API_KEY: ${{ secrets.TRIAGLY_API_KEY }}
    TRIAGLY_PROJECT_ID: ${{ secrets.TRIAGLY_PROJECT_ID }}
  run: |
    npx @triagly/cli export --format csv --output feedback-report.csv

Exit Codes

| Code | Meaning | |------|---------| | 0 | Success | | 1 | Auth error (missing or invalid key) | | 2 | Network or API error | | 3 | Invalid arguments | | 4 | Rate limited (1,000 req/min) |

Configuration File

~/.triagly/config.json (written by triagly login, updated by triagly use):

{
  "apiKey": "tsk_live_...",
  "projectId": "xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx",
  "baseUrl": "https://bssghvinezdawvupcyci.supabase.co/functions/v1"
}

Scopes

CLI-generated API keys have the feedback:read scope only. The CLI is entirely read-only — it cannot create, update, or delete feedback.