---

slug: polyglot-baseline-n7 title: Run polyglot §5 workingSetFit emitter against 5 more CLI tools date: 2026-05-24 branch: sama-v2-workingset-cross-repo-baseline pr_number: 33 merge_sha: 60056b1 status: lossy related_posts: [sama-v2-workingset-cross-repo-baseline]

Recovered opening (from prior-session conversation summary, paraphrased):

Goal: Run the polyglot §5 workingSetFit emitter against 5 additional popular open-source CLI tools (sharkdp/bat, sharkdp/fd, eza-community/eza, jesseduffield/lazygit, cli/cli) at pinned SHAs. Joining the existing dive + ripgrep measurements, the corpus becomes n=7 cross-repo datapoints (4 Rust + 3 Go) measured against the same [50, 500] LOC bounds. Publish a blog post with the distribution and the convergence question answered.

What is preserved from the original /goal: only the opening sentence above, recovered from the prior-session conversation summary at the start of this conversation. The full Done when clauses, Constraints (anti-fudge), and Load-bearing files sections did not survive summarization.

What landed: 7-datapoint cross-repo baseline with pinned SHAs throughout: tdd.md 80.00% (SAMA-disciplined), cli/gh 73.59%, sharkdp/fd 69.57%, lazygit 67.38%, eza 61.76%, ripgrep 54.00%, dive 52.17%, sharkdp/bat 46.27%. Range 27.32pp, mean 60.68%, sample stddev 10.13pp. The n=2 convergence (dive/ripgrep within 2pp) was confirmed as coincidence; 5 of 7 still cluster in [52%, 70%]. Hand-trace of bat (lowest measurement) included for /sama/v2 §0 auditability. See PR #33.