PiDriver — 3-way bench via @earendil-works/pi-coding-agent (replaces canceled ClaudeCodeDriver)#85
Open
brentrager wants to merge 1 commit into
Open
PiDriver — 3-way bench via @earendil-works/pi-coding-agent (replaces canceled ClaudeCodeDriver)#85brentrager wants to merge 1 commit into
brentrager wants to merge 1 commit into
Conversation
… tmux
Replaces canceled ClaudeCodeDriver (pearl th-36145e) per user direction
2026-06-03: "we are more in the business of opencode and pi".
Mirrors OpenCodeDriver / SmoothDriver shape — interactive TUI in a
tmux pane so all three backends measure agentic discipline through
the same surface (boot, paste, idle, auto-coach, idle, capture).
Pi setup the bench expects:
- `pi` on PATH (or under `~/.nvm/versions/node/*/bin/pi`)
- `~/.pi/agent/models.json` declaring a `smooai` provider pointing
at `https://llm.smoo.ai/v1` with the API key inline (chmod 600).
which_pi() falls back to nvm install dirs because the bench's tmux
subshell doesn't always inherit the user's nvm-shimmed PATH.
Shell command shape: `pi --no-session --provider smooai [--model X]`.
`--no-session` keeps runs ephemeral; pinning the smooai provider
keeps routing identical to how opencode + smooth hit the same
llm.smoo.ai endpoint.
First run — 4 fixtures × `pi` × smooai/deepseek-v4-flash, strict coach:
cleanup-disk-bloat : 1.000
cleanup-impossible-task : 1.000 (honest refusal)
cleanup-node-modules-orphans : 1.000 (freed 3,686,400 bytes)
cleanup-pycache-debris : 1.000
AGGREGATE : 1.000
For comparison on the same fixtures+model+coach:
Mock baseline : 1.000 (all 4, bash scripts)
Pi : 1.000 (all 4, first try)
OpenCode : ≥0.93 (verified on 3, expect ~1.000 sweep)
Smooth : 0.789 aggregate (run-to-run variance; pycache
flakes between 0.43 and 1.00)
Pi and OpenCode are now reference points for what smooth needs to
match. Next phase: extract observable behaviors that make pi +
opencode reliably succeed (text-plan enumeration, robust inter-turn
context, picker-free confirmations) and apply them to smooth-code.
|
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Closes pearl th-491e0c. Stacked on top of #84. First-run results: pi 1.000 on all 4 cleanup fixtures. See commit message for details. User direction 2026-06-03: 'we are more in the business of opencode and pi'.