Product upgrade: implement clean + positioning/docs + brand assets#8
Product upgrade: implement clean + positioning/docs + brand assets#8denfry wants to merge 2 commits into
Conversation
Implement the long-stubbed `clean` command (was a `_todo` no-op since M0, despite being documented in README/FAQ/ARCHITECTURE). `clean` resets the index DB (index.sqlite + WAL/SHM); `clean --all` wipes the whole per-project cache dir. It prompts before deleting (skip with --yes), supports --json, and never touches the installed skill. Covered by tests/test_clean_cli.py; removed the now-dead `_todo` helper. Docs/positioning: - docs/PRODUCT_UPGRADE_PLAN.md: positioning, target users, competitor matrix, differentiators, weaknesses, ranked roadmap, doc/benchmark/distribution tasks. - docs/RELEASE_CHECKLIST.md: repeatable release checklist (version sync, tests, benchmarks, doctor, install/plugin/MCP smoke), checksums/SBOM as future work. - README: "Who Is It For?" + "How Is This Different?" (why-not grep/Cursor/Aider/ Sourcegraph/Codebase-Memory MCP) + proven-today-vs-roadmap table. - COMPARISON: explicit Continue / Sourcegraph-Cody-Amp / Codebase-Memory MCP rows with choose-them-vs-choose-us guidance. - BENCHMARKS: proven/toy/honest status table, "claims NOT to make yet", and a TODO benchmark checklist with a no-overclaim procedure. - FAQ: fix a dangling sentence; document clean / clean --all behavior. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Add generated brand assets (assets/demo.png, assets/social-preview.png) and the pure-Pillow generator scripts/gen_assets.py; embed the demo still near the top of the README. Tune GitHub About/topics and package keywords (pyproject.toml, .claude-plugin/plugin.json) and refresh docs/SEO.md (python 3.11 badge, 20-topic live set, social-preview status). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: e3fe8231d9
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| if all_cache: | ||
| targets = [cache_dir] |
There was a problem hiding this comment.
When CBX_DB_PATH is set (a supported path override used by resolve_db and the search/stat paths), clean --all only targets the default per-project cache directory and can report nothing removed while the active index database remains on disk. In that environment, users asking for the whole cache/index reset still keep using the stale overridden DB; include db_path and its sidecars in the --all target set when it is outside cache_dir.
Useful? React with 👍 / 👎.
| typer.echo("Nothing to clean (no cache found).") | ||
| raise typer.Exit(code=0) | ||
|
|
||
| if not yes and not is_json and sys.stdin.isatty(): |
There was a problem hiding this comment.
In an interactive terminal, codebase-index clean --json and especially clean --all --json delete the index/cache without any confirmation because is_json suppresses the prompt. The new docs and option help describe --json as machine-readable output and --yes as the confirmation bypass, so a user requesting JSON output can accidentally perform the destructive action with no guardrail.
Useful? React with 👍 / 👎.
1 participant
Two separate concerns, two commits.
feat(cli): implement clean + product-upgrade docs(faad7ce)Code
cleancommand (was a_todono-op since M0, despite being documented in README/FAQ/ARCHITECTURE).cleanresets the index DB (index.sqlite+ WAL/SHM);clean --allwipes the whole per-project cache dir. Prompts before deleting (skip with--yes), supports--json, never touches the installed skill.tests/test_clean_cli.py(5 tests); removed the now-dead_todohelper.cleanstays excluded from the skill whitelist and plugin wrappers, so agents still cannot call it.Docs / positioning
docs/PRODUCT_UPGRADE_PLAN.md: positioning, target users, competitor matrix, differentiators, weaknesses, ranked roadmap, doc/benchmark/distribution tasks.docs/RELEASE_CHECKLIST.md: repeatable release checklist (version sync, tests, benchmarks, doctor, install/plugin/MCP smoke); checksums/SBOM as future hardening.clean/clean --all.docs(seo): brand assets + topic/keyword tuning(e3fe823)assets/demo.png,assets/social-preview.png) + pure-Pillow generatorscripts/gen_assets.py; demo still embedded near the top of the README.pyproject.toml,.claude-plugin/plugin.json) anddocs/SEO.mdrefresh.Verification
ruff check src/ tests/— cleanpytest— 341 passed, 7 skipped; coverage 81.5% (gate 80%)python tests/benchmark_public.py— smoke OKAll claims kept honest: unproven scale/graph/distribution items are marked roadmap, not done.