We read every piece of feedback, and take your input very seriously.
To see all available qualifiers, see our documentation.
docs: dashboard refonte — Compare tab, Prompt/File Explorer, Files constellation Reflect the reorganized dashboard: the Composition page (4 KPIs + 2 charts per section, Files constellation), the split Prompt Explorer / File Explorer with deep-links, the dedicated Compare tab (replacing the per-page before/after toggle), the extended global cross-filters, and the removed Prompts tab. Fix the stale references in Architecture / Codebase-Guide / CLI-Commands / Prompt-Categories.
docs(wiki): document the cost-by-content zone (v2.0.0) - CLI-Commands: new "Cost by content (composition)" section with by-output / by-context / by-file / by-task / impact; categorize --semantic / --audit / --provider azure; recipes. - Dashboard: the Composition page (input/output/context/files/tasks) and the global before/after compare mode. - Architecture: the additive composition CSVs in the data model; the new core modules (compose / context / tasks / tokenizer / embeddings / taxonomy_audit); analytics + dashboard module rows updated; thirteen categories. - Codebase-Guide: module map + data-model rows for the composition zone. - Home / Prompt-Categories: three classifiers (heuristic / semantic / LLM).
docs: link Prompt Categories from Home; English triggers Add the page to the Home documentation table (it was only in the sidebar). Render the per-category trigger list in English with a single "bilingual (FR+EN)" note, rather than mixing French triggers into an English page. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
docs: add Prompt Categories page Document the 13-label taxonomy: what each category represents, the heuristic decision pipeline (chrome stripping, notification + ack short-circuits, weighted scoring, tie-break, fallback), the weights and trigger words per category, how to read the distribution, observed complexity, and re-classification. Linked from the sidebar and the Dashboard page's Prompts row. Leads with a clear caveat that the rules are tuned on one developer's (non-representative) history and are the area where community contributions matter most. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>