chore: remove Glama integration#88
Conversation
Drops the Glama score badge from README.md and deletes the glama.json config file at the repo root. Glama's "Maintenance" sub-score is driven primarily by `issues responded to in the last N months`, which conflicts with this project's use of GitHub issues as a planned roadmap — the baggage tracker (#51) plus per-item issues are TODO entries, not unanswered support requests. The badge therefore mis-represents project health to anyone landing on the README, and the heuristic isn't tunable from our side. Removing rather than gaming the metric or adding a workflow note. PyPI / GitHub release / MCP registry remain the canonical distribution and discovery surfaces. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Codecov Report✅ All modified and coverable lines are covered by tests. 📢 Thoughts on this report? Let us know! |
cmeans
left a comment
There was a problem hiding this comment.
QA Round 1 — PASS (no findings)
Tight chore PR. Diff is exactly what the description says: 3 files / +4 / -5.
Verified
- README.md: drops the single
[]badge line; all other badges (PyPI, Python versions, License, Tests, Coverage, Downloads, per-installer, per-OS) preserved. - glama.json: deleted.
ls glama.json→ "No such file or directory". The file was minimal anyway ($schema+maintainers: ["cmeans"]); nothing else referenced it. - CHANGELOG.md: new
### Changedentry under## Unreleased, correctly slotted ABOVE## 0.5.2 (2026-05-01). Standard Keep a Changelog category. First PR after the v0.5.2 release introduces the section per project convention. Rationale text matches the PR body — the heuristic conflict between issue-as-roadmap and Glama's "issues responded to" metric is honestly stated. - Repo-wide
grep -in glamaturns up two residual hits, both intentional / accurate:.claude/settings.local.json:75— gitignored (git check-ignoreconfirms), local-only Claude settings; not part of the repo.scripts/sync-server-json.py:5— docstring notingserver.jsonis consumed by "the MCP Registry and Glama.ai". Glama still scansserver.jsonfrom the repo whether or not we maintain aglama.jsonor display a badge, so this comment is still factually correct. Not flagging.
Test plan checkboxes (all 3 pre-validated locally)
- CI green — 12/12 required incl. vdsm SUCCESS on
c342a9e. -
grep -in glama README.mdreturns nothing (exit 1, no match). -
glama.jsonno longer exists at repo root.
Local stack
uv run pytest→ 581 passed, 112 deselected, 96.19% coverage. (Same baseline as post-#85; this PR doesn't touch any code paths.)uv run ruff check,ruff format --check,mypy src/all clean.
Disposition
Ready for QA Signoff applied as the final act. Awaiting maintainer's QA Approved.
If you also delist on Glama's admin side (separate from this PR), the scripts/sync-server-json.py:5 docstring becomes the only remaining repo reference and would warrant a one-line drop of "and Glama.ai" — but that's a follow-up, not a blocker for this PR.
|
Applying Ready for QA Signoff — clean chore PR, no findings. Diff exactly matches the description: README badge line removed, glama.json deleted, CHANGELOG ## Unreleased ### Changed entry added with the rationale we discussed earlier. Local 581/96.19%, ruff/format/mypy clean, CI 12/12 green. Test-plan checkboxes 1-3 flipped (all pre-validated). Heads-up: if you also delist on Glama's admin side, scripts/sync-server-json.py:5 docstring should drop "and Glama.ai" as a follow-up — not blocking this PR. |
Summary
Drops the Glama score badge from
README.mdand deletes theglama.jsonconfig file at the repo root.Why
Glama's Maintenance sub-score is driven primarily by
issues responded to in the last N months, which conflicts with this project's use of GitHub issues as a planned roadmap. The baggage tracker (#51) plus per-item issues (#33–47, #75/76, etc.) are TODO entries — work we know we want to do — not unanswered support requests. Glama can't tell the difference, so the score reads as "0 of 33 issues responded to" → C grade on Maintenance, even though every one of those issues is intentionally tracked and being worked through (we just merged six of them in v0.5.2 alone).The badge therefore mis-represents project health to anyone landing on the README, and the heuristic isn't tunable from our side. Removing rather than gaming the metric (e.g. bulk-commenting "tracked elsewhere" on every open issue) or adding a workflow note that wouldn't fix the perception for casual readers.
What's left
PyPI, GitHub release, and the MCP registry remain the canonical distribution and discovery surfaces. The README still carries the standard PyPI / Python-versions / License / Tests / Coverage / Downloads badges plus the per-installer and per-OS download breakdowns from
cmeans/pypi-winnow-downloads.Files changed
README.md— drops the single[]lineglama.json— deleted (was just{ "$schema": "...", "maintainers": ["cmeans"] })CHANGELOG.md—### Changedentry under## Unreleasedexplaining the rationaleTest plan
grep -in glama README.mdreturns nothingglama.jsonno longer exists at repo root🤖 Generated with Claude Code