NEVER GENERATE GPD COMMANDS WITHOUT RUNNING gpd --help OR SUBCOMMAND --help FIRST
Always use --help to discover current flags and subcommands. The gpd CLI evolves — never assume flags from memory.
Command discovery
Always use --help to confirm commands and flags.
gpd --help
gpd publish --help
gpd monetization --help
Flag conventions
Use explicit long flags (for example: --package, --track, --status).
No interactive prompts; destructive operations require --confirm.
Use --all when the user wants all pages.
Output formats
Default output is minified JSON.
Use --pretty for readable JSON during debugging.
Authentication and defaults
Service account auth via GPD_SERVICE_ACCOUNT_KEY is required.
Validate access for a package:
gpd auth check --package com.example.app
Safety
Use --dry-run when available before destructive operations.
Prefer edit lifecycle (gpd publish edit create) for multi-step publishing.
Anti-Patterns
Don't generate gpd commands from memory. The gpd CLI evolves — flags and subcommands change between versions. Always run gpd <subcommand> --help before generating a command; the Iron Law exists for this reason.
Don't omit --confirm on destructive operations. Commands like gpd publish halt and gpd publish rollback require --confirm; omitting it will either fail or prompt interactively, breaking automation.
Don't use short flags in scripts. Always use explicit long flags (--package, --track, --status) to ensure scripts are readable and resilient to positional changes.
Verify
After running any gpd command:
Check that the exit code is 0: echo $? — a non-zero exit means the operation failed even if no error was printed.
Validate auth before running a sequence: gpd auth check --package com.example.app — a missing or expired GPD_SERVICE_ACCOUNT_KEY will cause silent failures mid-sequence.
Documentation Sources
Source
How to Access
Purpose
gpd CLI help
gpd --help, gpd <command> --help
Discover current flags, subcommands, and options
gpd subcommand help
gpd publish --help, gpd monetization --help
Command-specific flags and usage
1---2name: gpd-cli-usage3description: GPD CLI Usage4---56# GPD CLI Usage78## Iron Law910**NEVER GENERATE GPD COMMANDS WITHOUT RUNNING `gpd --help` OR SUBCOMMAND `--help` FIRST**1112Always use `--help` to discover current flags and subcommands. The gpd CLI evolves — never assume flags from memory.1314## Command discovery15- Always use `--help` to confirm commands and flags.16 - `gpd --help`17 - `gpd publish --help`18 - `gpd monetization --help`1920## Flag conventions21- Use explicit long flags (for example: `--package`, `--track`, `--status`).22- No interactive prompts; destructive operations require `--confirm`.23- Use `--all` when the user wants all pages.2425## Output formats26- Default output is minified JSON.27- Use `--pretty` for readable JSON during debugging.2829## Authentication and defaults30- Service account auth via `GPD_SERVICE_ACCOUNT_KEY` is required.31- Validate access for a package:32 - `gpd auth check --package com.example.app`3334## Safety35- Use `--dry-run` when available before destructive operations.36- Prefer edit lifecycle (`gpd publish edit create`) for multi-step publishing.3738## Anti-Patterns3940- **Don't generate gpd commands from memory.** The gpd CLI evolves — flags and subcommands change between versions. Always run `gpd <subcommand> --help` before generating a command; the Iron Law exists for this reason.41- **Don't omit `--confirm` on destructive operations.** Commands like `gpd publish halt` and `gpd publish rollback` require `--confirm`; omitting it will either fail or prompt interactively, breaking automation.42- **Don't use short flags in scripts.** Always use explicit long flags (`--package`, `--track`, `--status`) to ensure scripts are readable and resilient to positional changes.4344## Verify4546After running any gpd command:47- Check that the exit code is `0`: `echo $?` — a non-zero exit means the operation failed even if no error was printed.48- Validate auth before running a sequence: `gpd auth check --package com.example.app` — a missing or expired `GPD_SERVICE_ACCOUNT_KEY` will cause silent failures mid-sequence.4950## Documentation Sources5152| Source | How to Access | Purpose |53|--------|--------------|---------|54| gpd CLI help | `gpd --help`, `gpd <command> --help` | Discover current flags, subcommands, and options |55| gpd subcommand help | `gpd publish --help`, `gpd monetization --help` | Command-specific flags and usage |
Run npx skillmds@latest add kumaran-is/gpd-cli-usage in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.
GPD CLI Usage It is listed under Coding & Dev Tools on SkillMD.
This skill has not completed SkillMD's automated safety review yet. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.
This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.
Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.
kumaran-is (@kumaran-is) published this skill. Their other Agent Skills are listed on their SkillMD profile.