Add a set of Agent Skills bundled with the repository so that AI coding agents (Claude Code, and any tool supporting the open Agent Skills standard) can drive clockify-cli reliably for common time-tracking workflows.
Agent Skills are an open, cross-tool standard: a skill is a directory containing a SKILL.md file with YAML frontmatter (name, description, optional allowed-tools, etc.) plus the markdown instructions an agent follows. Shipping skills in-repo means users get curated,
tested guidance instead of agents guessing at flags and subcommands.
clockify-cli is a natural fit for agent automation - starting/stopping entries, reporting, and editing are repetitive tasks users already script. Today an agent has to infer command syntax from --help output or the docs site, which is error-prone. A small set of vetted skills would make the CLI a first-class tool for agents.
Proposal
Add a skills/directory at the repo root, with one skill per workflow. Each skill is a folder with a SKILL.md:
skills/
├── clockify-log-time/SKILL.md # start/stop/clone entries
├── clockify-report/SKILL.md # report over date ranges
├── clockify-edit-entry/SKILL.md # edit existing entries
└── clockify-setup/SKILL.md # init/config management
Or some variation of the above.
Each skill body would cover the relevant subcommands, the auto-filter options (last-month, this-month), how non-interactive flags map to the interactive prompts, and how to resolve workspace/project names.
Add a set of Agent Skills bundled with the repository so that AI coding agents (Claude Code, and any tool supporting the open Agent Skills standard) can drive clockify-cli reliably for common time-tracking workflows.
Agent Skills are an open, cross-tool standard: a skill is a directory containing a SKILL.md file with YAML frontmatter (name, description, optional allowed-tools, etc.) plus the markdown instructions an agent follows. Shipping skills in-repo means users get curated,
tested guidance instead of agents guessing at flags and subcommands.
clockify-cli is a natural fit for agent automation - starting/stopping entries, reporting, and editing are repetitive tasks users already script. Today an agent has to infer command syntax from --help output or the docs site, which is error-prone. A small set of vetted skills would make the CLI a first-class tool for agents.
Proposal
Add a skills/directory at the repo root, with one skill per workflow. Each skill is a folder with a SKILL.md:
skills/
├── clockify-log-time/SKILL.md # start/stop/clone entries
├── clockify-report/SKILL.md # report over date ranges
├── clockify-edit-entry/SKILL.md # edit existing entries
└── clockify-setup/SKILL.md # init/config management
Or some variation of the above.
Each skill body would cover the relevant subcommands, the auto-filter options (last-month, this-month), how non-interactive flags map to the interactive prompts, and how to resolve workspace/project names.