docs: plain-prose pass over README and published docs

[lesnik512]

Prose pass over README.md and every page mkdocs publishes (docs/, excluding docs/agents and docs/adr). Code blocks, inline code, link targets, admonition syntax and headings are unchanged, so no anchors moved.

What changed

• Em and en dashes in prose: 207 to 0. They became colons, commas, semicolons, parentheses or rewrites. Numeric ranges (C1-C3, 5-11%) use hyphens. Most were "[Link] — description" in See also lists and table cells. Dashes left in code-block comments are untouched.
• Bold run-in labels: 116 to 0. This covers "Problem." recipe openers, the "Caught by:" lines in good-and-bad-practices, "Fix:" and "Direct miss." in troubleshooting, bold numbered Fix/Cause items, glossary bullets in errors-and-exceptions, per-provider "X →" labels in the migration guides, and the thread-safety bullets in design-decisions. Bold spans overall went from 266 to 58. The rest are inline emphasis on an ordering constraint (before, after, must), a term being defined, or the ratio cells in performance.md.
• Not-X-but-Y contrasts and "rather than"/"instead of": rewritten where the negative half added only weight, e.g. "not a coordination tool", "a design decision, not a gap", "is application code, not a framework feature", "isn't about features so much as". Kept where the negative half corrects something a reader from FastAPI, dishka, that-depends or dependency-injector would assume (e.g. "keyed by provider reference, not attribute name", "not a hook").
• Also cut: staged openers ("modern-di isn't the only way..."), repeated closers, filler ("simply", "just", "genuinely"), and "no longer"/"as of" wording outside the migration guides and the performance history. Version facts stay (3.1, removal in 4.0).
• Title Case: README "Quick Start" becomes "Quick start".

Other change

• docs/assets/social-card.png is replaced with the card regenerated from the current description in docs: sync social-card taglines with repo descriptions .github#125.

Verification

• just lint-ci: pass
• just test-ci: 548 passed, 100% coverage
• uvx --with-requirements docs/requirements.txt mkdocs build --strict: pass (anchor validation on)

Not changed

• The markdown_description in mkdocs.yml ("Powerful ...") mirrors the pyproject description, which this pass does not touch.
• Benchmark tables and footnotes in performance.md are generated by just bench-report. Only the hand-written prose around them changed.

Fact fix

docs/recipes/sqlalchemy.md said "Three providers, three scopes"; the engine is APP and both the session and repositories are REQUEST, so it now says two scopes.

[github-actions]

Benchmark

Details

Benchmark suite	Current: 3d2a3ec	Previous: 3bb845c	Ratio
benchmarks/test_guard_by_type.py::test_g16_resolve_by_type	3748966.7847559517 iter/sec (stddev: 6.536767954885753e-9)	3552595.6276174076 iter/sec (stddev: 1.8017495134377973e-8)	0.95
benchmarks/test_guard_by_type.py::test_g17_resolve_by_type_large_registry	3744696.1995632937 iter/sec (stddev: 8.11036842585726e-9)	3525743.000987044 iter/sec (stddev: 1.9520895722614712e-8)	0.94
benchmarks/test_guard_cold.py::test_g8_cold_first_resolve	15262.283863453004 iter/sec (stddev: 0.000056144324850628354)	15442.399035232269 iter/sec (stddev: 0.0000504562631484814)	1.01
benchmarks/test_guard_cold.py::test_g8b_cold_first_resolve_cached	13002.56170014031 iter/sec (stddev: 0.00012105475718252928)	12987.261259255825 iter/sec (stddev: 0.00013444409804140028)	1.00
benchmarks/test_guard_concurrency.py::test_g14_concurrent_cached_hit[1]	455.3955817409212 iter/sec (stddev: 0.000050486138046866206)	370.149961520189 iter/sec (stddev: 0.00006432698950621452)	0.81
benchmarks/test_guard_concurrency.py::test_g14_concurrent_cached_hit[2]	419.09836567790626 iter/sec (stddev: 0.000039590248208970655)	353.9950395466471 iter/sec (stddev: 0.00004818188411818694)	0.84
benchmarks/test_guard_concurrency.py::test_g14_concurrent_cached_hit[4]	377.1765699010654 iter/sec (stddev: 0.00011017185804252727)	325.7458820366416 iter/sec (stddev: 0.00005350777640166116)	0.86
benchmarks/test_guard_concurrency.py::test_g15_concurrent_first_resolve[1]	1509.1429729233507 iter/sec (stddev: 0.00018523510489365963)	1591.09633100402 iter/sec (stddev: 0.00019877711335045307)	1.05
benchmarks/test_guard_concurrency.py::test_g15_concurrent_first_resolve[2]	1200.9439179018766 iter/sec (stddev: 0.00024118857648077488)	1251.96977882641 iter/sec (stddev: 0.00025669199194511646)	1.04
benchmarks/test_guard_concurrency.py::test_g15_concurrent_first_resolve[4]	872.6382699702993 iter/sec (stddev: 0.0002500576515871365)	948.6776536343597 iter/sec (stddev: 0.00019771596542946792)	1.09
benchmarks/test_guard_concurrency.py::test_g15b_concurrent_first_resolve_sibling_children[1]	1264.863049956556 iter/sec (stddev: 0.0012603372528518645)	1341.9049449995864 iter/sec (stddev: 0.0013131786726602232)	1.06
benchmarks/test_guard_concurrency.py::test_g15b_concurrent_first_resolve_sibling_children[2]	1121.4517477002366 iter/sec (stddev: 0.00021032087247726763)	1157.5662043227023 iter/sec (stddev: 0.00020992030300014155)	1.03
benchmarks/test_guard_concurrency.py::test_g15b_concurrent_first_resolve_sibling_children[4]	719.0375940654415 iter/sec (stddev: 0.0002691176420246703)	771.876181875189 iter/sec (stddev: 0.00023582272008191215)	1.07
benchmarks/test_guard_lifecycle.py::test_g6_build_child_container	817861.2723259545 iter/sec (stddev: 7.574284011913898e-8)	799413.3425239581 iter/sec (stddev: 4.6161785509303694e-8)	0.98
benchmarks/test_guard_lifecycle.py::test_g6b_build_child_container_auto_scope	785911.0048580153 iter/sec (stddev: 3.95400831295033e-8)	741469.4129403305 iter/sec (stddev: 8.216022529812231e-8)	0.94
benchmarks/test_guard_lifecycle.py::test_g7_request_lifecycle_batch	2423.122423645699 iter/sec (stddev: 0.000013429193034333899)	2418.1335157858957 iter/sec (stddev: 0.000011054809907129041)	1.00
benchmarks/test_guard_lifecycle.py::test_g7c_event_loop_floor_control	56207.67718364238 iter/sec (stddev: 0.000002660081497777069)	55362.81582342883 iter/sec (stddev: 0.0000037110526828629025)	0.98
benchmarks/test_guard_lifecycle.py::test_g13_teardown_at_scale	45445.524114963955 iter/sec (stddev: 0.0000020907397910757774)	35058.66382477332 iter/sec (stddev: 0.00000686006290349675)	0.77
benchmarks/test_guard_lifecycle.py::test_g13b_teardown_at_scale_async_no_finalizers	578.5642523083935 iter/sec (stddev: 0.000028699674660190535)	552.0029132868616 iter/sec (stddev: 0.000024694912850732793)	0.95
benchmarks/test_guard_resolve.py::test_g1_transient_resolve	2226720.3209929727 iter/sec (stddev: 5.478080783414517e-8)	2164661.000023137 iter/sec (stddev: 2.9672373641740744e-8)	0.97
benchmarks/test_guard_resolve.py::test_g2_cached_resolve	3594999.054309258 iter/sec (stddev: 1.133302239101231e-8)	3478358.9895358584 iter/sec (stddev: 2.2492268429006365e-8)	0.97
benchmarks/test_guard_resolve.py::test_g3_deep_chain	750179.0349159556 iter/sec (stddev: 6.482694244770068e-8)	734894.0596930252 iter/sec (stddev: 4.385126547025142e-8)	0.98
benchmarks/test_guard_resolve.py::test_g4_wide_resolve	465230.6067828385 iter/sec (stddev: 3.4785942847911485e-7)	448788.90051667934 iter/sec (stddev: 3.574176524885024e-7)	0.96
benchmarks/test_guard_resolve.py::test_g5_cross_scope	1992556.010003286 iter/sec (stddev: 3.400860028373226e-8)	1993989.5173966833 iter/sec (stddev: 3.37719590410727e-8)	1.00
benchmarks/test_guard_resolve.py::test_g9_context_resolve	1050736.7963374218 iter/sec (stddev: 1.7683787190997395e-7)	1040260.310581399 iter/sec (stddev: 1.6805736399783922e-7)	0.99
benchmarks/test_guard_resolve.py::test_g12_override_active_resolve	711167.0436069932 iter/sec (stddev: 9.366771760959301e-8)	738595.5583310551 iter/sec (stddev: 4.479275195662805e-8)	1.04
benchmarks/test_guard_resolve.py::test_g18_alias_hop	2298485.617464901 iter/sec (stddev: 1.1021860597509296e-8)	1972756.1585418966 iter/sec (stddev: 2.1135635307380664e-8)	0.86
benchmarks/test_guard_validate.py::test_g10_validate_deep_chain	23190.658270988817 iter/sec (stddev: 0.000022741002005993537)	23479.450624332232 iter/sec (stddev: 0.000023210661423516627)	1.01
benchmarks/test_guard_validate.py::test_g11_validate_wide	14140.246833074865 iter/sec (stddev: 0.000027656583936823687)	14154.42798938568 iter/sec (stddev: 0.000028768193092020475)	1.00

This comment was automatically generated by workflow using github-action-benchmark.