@brettinternet/pi-progress

Compact passive activity and verification progress for Pi

Packages

Package details

extension

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

npm repohome report
$ pi install npm:@brettinternet/pi-progress
Package
@brettinternet/pi-progress
Version
0.1.6
Published
Sep 29, 2026
Downloads
500/mo · 17/wk
Author
brettinternet
License
MIT
Types
extension
Size
61.3 KB
Dependencies
1 dependency · 2 peers
Pi manifest JSON
{
  "extensions": [
    "./index.ts"
  ]
}

Security note

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

README

pi-progress

Show what the agent is doing below the editor.

progress 25m · agents 42m · current: Updating the implementation inferred · ● edit src/index.ts · ✓ bun test
 touched src/index.ts · test/index.test.ts
pi install npm:@brettinternet/pi-progress
Segment Meaning
progress 25m Work time
agents 42m Summed subagent runtime; hidden until a subagent reports
current: … inferred Model-inferred step; off until you set a model
● edit src/index.ts Running tool
✓ bun test Finished check
touched … Files edited or written successfully

These show activity, not proof the result is correct. Inferred claims that something was "verified" are dropped without discarding other progress labels.

Commands

/progress status                        show inference state
/progress steps                         toggle recent history   (Alt+G)
/progress steps recent                  show the last eight steps
/progress steps all                     show full history       (Alt+Shift+G)
/progress model                         show the inference model
/progress model openai/gpt-5-nano:low   enable inference
/progress model off                     disable inference

Configuration

// ~/.pi/agent/pi-progress.jsonc
{
  "model": null,           // "provider/model[:effort]" enables inference
  "maxInputChars": 12000,
  "maxTokens": 180,
  "timeoutMs": 15000
}

The inference model gets a bounded, redacted digest. It never sees reasoning, tool output, file contents, diffs, environment data, credentials, or the full transcript.