---

name: cli-and-project-ops description: "Guides the PandasAI pai CLI, login and dataset-create commands, API-key validation, project-root behavior, and lightweight repo maintenance operations." disable-model-invocation: true metadata: disco-role: operating license: NOASSERTION

CLI and Project Ops

Use this sub-skill when a task involves the pai command, PandaBI API-key login, guided CLI dataset creation, project-root/datasets layout, or maintainer commands for a PandasAI checkout.

Fast route

1. Verify the CLI is importable and visible:

   pai --help
   pai dataset --help
   

2. For login, validate the key format before writing .env:

   pai login PAI-xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx
   

3. For interactive dataset creation, run from the intended project root because .env and datasets/ paths are resolved relative to project-root discovery.

4. For programmatic schema creation or debugging, route to ../semantic-layer/SKILL.md.

5. For source-code edits or tests, use the maintainer notes and choose focused tests instead of all extension targets unless the user needs them.

Read next

• references/cli-reference.md for command syntax, prompts, and validation behavior.
• references/project-layout.md for project root, .env, and datasets/ placement.
• references/maintainer-notes.md for Poetry, Makefile, and focused test-selection guidance.
• references/troubleshooting.md for missing CLI dependencies, invalid keys, interactive prompt issues, and existing datasets.
• scripts/pai_cli_smoke.py for a safe no-credential CLI diagnostic.

Boundaries

• Route pai.create, pai.load, schemas, views, and transformations to ../semantic-layer/SKILL.md.
• Route LLM/provider setup and .chat() failures to ../conversational-analysis/SKILL.md.
• Route sandbox choices to ../sandbox-and-security/SKILL.md.

Safe validation

python sub-skills/cli-and-project-ops/scripts/pai_cli_smoke.py --show-help
python sub-skills/cli-and-project-ops/scripts/pai_cli_smoke.py --check-api-key PAI-59ca2c4a-7998-4195-81d1-5c597f998867
python sub-skills/cli-and-project-ops/scripts/pai_cli_smoke.py --isolated-login-smoke

The smoke uses Click's isolated runner and a fake valid-format key. It never uses real credentials.

Common gotchas

• CLI import can fail if click is missing from the active environment.
• pai login writes or updates PANDABI_API_KEY in .env under the detected project root and preserves unrelated variables.
• pai dataset create is interactive. Automate it only with an isolated test runner or a deliberate input stream.
• Dataset path validation is the same lowercase hyphenated organization/dataset contract used by the semantic layer.
• Full extension tests can install many optional packages; prefer focused tests when validating a small source change.