Beta groups from apple.testflight.groups (created when missing, never changed), testers from a CSV in .storebound/
(masked in logs, confirmation before Apple emails invitations, --remove / --invite), What to Test per locale from the
App Store listing (whatToTest or whatsNew), Test Information and Beta App Review details, builds by exact number to
groups, Beta App Review submission behind --submit-review and a confirmation. Pure planners in cli/lib/testflight.ts,
shared CSV / table / JSON:API helpers. MCP: apple_testflight_status, _groups, _what_to_test, _info, _distribute
(internal groups only). Skill testflight.
- Apple sizes accepted portrait or landscape; Play screenshots any accepted size (320-3840 px, at most 2:1);
Play rejects 16-bit PNGs; feature graphic exactly 1024 x 500
- Required sets per language (Apple only with an App Store locale, Play only with a Play code)
- shots check scans A/B variant folders only with --variants
- validate adds the local renders checks: store errors are blockers, overflow flags warnings (--no-local skips)
- Tablet frames default to the 13" iPad screen (1032 x 1352 @2x below a 24 px bar) when apple.ipad is set
- shots.capture.cropTop takes a number or { phone, tablet }; default is the frame's bar share per device class
- capture warns when a cropped frame does not match the frame aspect
- apple screenshots uploads ipad13 as APP_IPAD_PRO_3GEN_129 only with apple.ipad; validate requires it then
- ios-sim: one warm-up launch after a cold boot (ios.warmupMs, default 8000) so the first frame is not blank
simctl / adb / ffmpeg argv builders in devices.ts; capture.ts runs them with
per-shot progress, --dry, clear errors when xcrun / adb / ffmpeg are missing,
status bar restore, and shuts down only a simulator it booted (--keep-booted).
Salvaged from wip/shots-pipeline. Pure modules (spec, devices, report, video,
check), variant support in render / lib, 13" iPad size and 3:4 template layout.
Not all of it is wired into the CLI yet.
- storebound mcp: stdio MCP server (spec 2025-11-25, tools only), one tool per command, write tools
dry_run by default, readOnly/destructive/idempotent/openWorld annotations; submit, release,
rollout, cancel and delete are not exposed. Own JSON-RPC subset, no SDK, zero runtime deps.
- storebound install --agent claude|codex|cursor|copilot|windsurf|gemini|opencode: copies the skills
and adds the storebound MCP entry (.mcp.json, .codex/config.toml, .cursor/mcp.json, .vscode/mcp.json,
Windsurf mcp_config.json, .gemini/settings.json, opencode.json). Idempotent, keeps other entries, --dry.
- validate: read-only App Store Connect + Play report with ordered blockers (exit 1)
- apple release --phased on|pause|resume|complete and --manual-release (asks first)
- apple status --wait-review (exit 0 approved, 1 rejected, 7 timeout), apple next-build, apple encryption
- play rollout --fraction|--halt|--resume|--complete, play next-version-code
- check --aso, whats-new --since <tag>, privacy-manifest --ios <path>, version
- Apple IAP description limit is 45 characters (App Store Connect Help), not 55
- config: apple.usesNonExemptEncryption, apple.screenshotTypes, ads.appAdsTxt
