@czottmann/pi-whoami

pi extension that adds a `whoami` tool reporting the active provider, model, and thinking level.

Packages

Package details

extension

Install @czottmann/pi-whoami from npm and Pi will load the resources declared by the package manifest.

npm repohome report
$ pi install npm:@czottmann/pi-whoami
Package
@czottmann/pi-whoami
Version
1.0.0
Published
Aug 16, 2026
Downloads
198/mo · 21/wk
Author
czottmann
License
MIT
Types
extension
Size
8.6 KB
Dependencies
0 dependencies · 1 peer
Pi manifest JSON
{
  "extensions": [
    "./extensions/whoami.ts"
  ]
}

Security note

Pi packages can execute code and influence agent behavior. Review the source before installing third-party packages.

README

@czottmann/pi-whoami

A pi extension that adds a whoami tool. The tool reports the active provider, model, and thinking level of the current pi session.

Install

pi install npm:@czottmann/pi-whoami

From a local checkout:

pi install .

For one run from a local checkout:

pi -e ./extensions/whoami.ts

Usage

The extension registers one tool: whoami. It takes no arguments. Ask the agent something like "who am I running on?" and it can call the tool.

Example output:

{
  "provider": "openrouter",
  "model": "google/gemma-3-27b-it",
  "modelName": "Gemma 3 27B",
  "thinkingLevel": "high",
  "providerThinkingLevel": "high",
  "projectDir": "/Users/you/Code/my-project"
}

Fields

Field Description
provider The active provider ID, for example openai-codex or openrouter.
model The active model ID, for example glm-5.3.
modelName The human-readable model name. Equal to the model ID when the catalog sets no name.
thinkingLevel The pi thinking level: off, minimal, low, medium, high, xhigh, or max.
providerThinkingLevel The thinking value that pi sent to the provider. See below.
projectDir The directory where pi started the session.

providerThinkingLevel

Pi maps its thinking levels to values that the provider API accepts. The result depends on the provider and model:

  • Most providers get a level string, for example high or xhigh. Pi can clamp levels. For example, a model without max support can receive high.
  • Budget-based providers (older Anthropic models, some Google models) get a token budget. In this case the field holds a number.
  • The field is null when the provider received no thinking directive, for example when thinking is off.

The extension reads this value from the outgoing provider request. Before the first request of a session, it falls back to the model's thinkingLevelMap.

Development

npm run check
npm run build
pi -e . --provider nebius-token-factory

Publishing

GitHub Actions publishes the package to npm when a GitHub Release is published. Releases are tagged vX.Y.Z, where X.Y.Z must match package.json exactly (for example tag v1.0.0 for version 1.0.0).

The workflow uses npm Trusted Publishing, so it does not need an npm token secret. Configure this package on npm with this repository and workflow file (.github/workflows/publish.yml). The workflow builds the package, runs npm run check, and publishes with npm provenance.

Author

Carlo Zottmann, carlo@zottmann.dev