openclaw/byr-cli

BYR CLI Skill

When To Use

Use this skill when a task needs any BYR operation via CLI:

• authenticate/check auth state
• search torrents with filters
• browse latest torrents with filters
• inspect torrent details
• plan or execute torrent downloads
• fetch BYR metadata and user info
• run local diagnostics before live calls

Boundaries

• Work only through the byr binary.
• Prefer --json for machine-readable output.
• Do not infer missing IDs/paths or silently mutate files.
• Keep read-only commands non-destructive.

Auth Notes

• Support both cookie formats in auth import-cookie:
  • uid=...; pass=...
  • session_id=...; auth_token=... (optional refresh_token=...)
• Browser import:
  • chrome (macOS path/decrypt flow)
  • safari best effort with manual fallback
• Always check status before live operations:
  • byr auth status --verify --json

Commands (JSON First)

Read-only:

• byr check --json
• byr whoami --json
• byr doctor [--verify] --json
• byr browse [--limit <n>] [--category <alias|id>] [--incldead <alias|id>] [--spstate <alias|id>] [--bookmarked <alias|id>] [--page <n>] --json
• byr search --query "<text>" --limit <n> --json
• byr search --imdb <tt-id> [--category <alias|id>] [--spstate <alias|id>] --json
• byr get --id <torrent-id> --json
• byr user info --json
• byr meta categories --json
• byr meta levels --json
• byr auth status [--verify] --json
• byr auth import-cookie --cookie "<cookie-header>" --json
• byr auth import-cookie --from-browser <chrome|safari> [--profile <name>] --json
• byr auth logout --json

Write side effect:

• Dry run first: byr download --id <torrent-id> --output <path> --dry-run --json
• Actual write: byr download --id <torrent-id> --output <path> --json

Search/Browse Semantics

• search and browse return paged list data.
• JSON fields:
  • matchedTotal: estimated total hits inferred from BYR pagination range blocks.
  • returned: number of items returned in current payload.
  • total: backward-compatible alias of returned.
• If --page is omitted, list commands auto-fetch subsequent pages until --limit is reached.
• If --page is provided, only that page is fetched.

Side-Effect Policy

Before non-dry-run download:

1. verify --id and --output are explicit
2. run dry-run and inspect sourceUrl/fileName
3. confirm intent for the output path

If parameters are missing, ask for explicit values.

Error handling

• Surface error.code and error.message.
• For E_ARG_*: request corrected flags/arguments.
• For E_AUTH_*: re-auth guidance (auth import-cookie or credential refresh).
• For E_NOT_FOUND_*: request different query/torrent ID.
• For E_UPSTREAM_*: suggest retry and capture command/context.

Response Style

• Keep result summaries short.
• Include key fields for search/get: id, title, size, seeders, leechers.
• For list commands include both matchedTotal and returned when present.
• Include key fields for download: outputPath, sourceUrl, dryRun, bytesWritten.