docs(decision): the @inject asymmetry is inherent

[lesnik512]

Closes the blessed-ready audit's D5 (@inject) thread by ruling the 4-vs-8 decorator-free/@inject split inherent — not an adapter gap. Full rationale + seam table: planning/decisions/2026-07-25-inject-asymmetry-inherent.md.

FromDI goes decorator-free only where the framework evaluates a parameter default as a provider (fastapi.Depends, litestar Provide, faststream.Depends, TaskiqDepends). The eight @inject integrations expose no such per-parameter seam (plain callables / fixed signatures), so @inject is required — an adapter cannot manufacture a seam the framework lacks. Every one is a 1, none a 0 (no unused seam). aiogram is the close call — its name-based data injection isn't provider-evaluation, so it stays a 1.

Positioning note kept: no @provide ever, and no @inject in the four biggest integrations — not "decorator-free" unqualified.

Consistent with the D3-inherent and exec-declined precedents; concrete revisit trigger (a Category-8 framework gains a DI hook). Adds the forward pointer from the audit's §4 backlog. Docs-only; check-planning + lint-ci green.

🤖 Generated with Claude Code

[github-actions]

Benchmark

Details

Benchmark suite	Current: 61e5e6d	Previous: 0cb20f1	Ratio
benchmarks/test_guard_cold.py::test_g8_cold_first_resolve	22728.01368184558 iter/sec (stddev: 0.00017300452739533794)	19363.01269293753 iter/sec (stddev: 0.0001737568722598439)	0.85
benchmarks/test_guard_concurrency.py::test_g14_concurrent_cached_hit[1]	469.3113666110011 iter/sec (stddev: 0.00009209763983747628)	291.29401500564364 iter/sec (stddev: 0.0000984051347273799)	0.62
benchmarks/test_guard_concurrency.py::test_g14_concurrent_cached_hit[2]	395.2085123111992 iter/sec (stddev: 0.0004154034941035602)	284.14059090572823 iter/sec (stddev: 0.00010971943220177788)	0.72
benchmarks/test_guard_concurrency.py::test_g14_concurrent_cached_hit[4]	418.7956265768672 iter/sec (stddev: 0.00006501005413083908)	264.5194577976377 iter/sec (stddev: 0.000044426493980133685)	0.63
benchmarks/test_guard_concurrency.py::test_g15_concurrent_first_resolve[1]	2101.9335108339055 iter/sec (stddev: 0.00031015519873394754)	1599.0893079765003 iter/sec (stddev: 0.0004914871064751759)	0.76
benchmarks/test_guard_concurrency.py::test_g15_concurrent_first_resolve[2]	1344.2278369397995 iter/sec (stddev: 0.0012063417852012923)	1087.5940350796577 iter/sec (stddev: 0.0013759420395216704)	0.81
benchmarks/test_guard_concurrency.py::test_g15_concurrent_first_resolve[4]	1088.6459345656597 iter/sec (stddev: 0.0003900597908673867)	865.3246375709942 iter/sec (stddev: 0.0006609327848643371)	0.79
benchmarks/test_guard_lifecycle.py::test_g6_build_child_container	567678.5681749944 iter/sec (stddev: 0.000029821198292496153)	449000.49034032243 iter/sec (stddev: 0.00004307837627762919)	0.79
benchmarks/test_guard_lifecycle.py::test_g6b_build_child_container_auto_scope	580568.6233291627 iter/sec (stddev: 0.000009259549734184697)	448208.1747958087 iter/sec (stddev: 0.000010305694696287827)	0.77
benchmarks/test_guard_lifecycle.py::test_g7_request_lifecycle	89045.97721602974 iter/sec (stddev: 0.000012268558487896937)	66050.37034553982 iter/sec (stddev: 0.000014127844918670553)	0.74
benchmarks/test_guard_lifecycle.py::test_g13_teardown_at_scale	52651.42350422114 iter/sec (stddev: 0.000018619066312879198)	41977.226833343266 iter/sec (stddev: 0.000020758288904301344)	0.80
benchmarks/test_guard_resolve.py::test_g1_transient_resolve	1726695.8708554003 iter/sec (stddev: 3.041646067033082e-7)	1182204.5907995403 iter/sec (stddev: 3.6611020650631277e-7)	0.68
benchmarks/test_guard_resolve.py::test_g2_cached_resolve	3551600.465229264 iter/sec (stddev: 5.8995395975809375e-8)	2481935.5415722583 iter/sec (stddev: 6.635869092266185e-8)	0.70
benchmarks/test_guard_resolve.py::test_g3_deep_chain	653064.9616417523 iter/sec (stddev: 8.72813798223193e-7)	480328.34348906146 iter/sec (stddev: 5.844070393569079e-7)	0.74
benchmarks/test_guard_resolve.py::test_g4_wide_resolve	414943.8827648377 iter/sec (stddev: 7.091962031804844e-7)	303343.11451952846 iter/sec (stddev: 6.581846470056526e-7)	0.73
benchmarks/test_guard_resolve.py::test_g5_cross_scope	1432200.2421315783 iter/sec (stddev: 4.2592807354277765e-7)	1038272.2338780607 iter/sec (stddev: 3.821044364902488e-7)	0.72
benchmarks/test_guard_resolve.py::test_g9_context_resolve	802673.546765506 iter/sec (stddev: 3.286076758768385e-7)	550106.0237511761 iter/sec (stddev: 4.951488903409311e-7)	0.69
benchmarks/test_guard_resolve.py::test_g12_override_active_resolve	581345.4618448393 iter/sec (stddev: 3.6853716267577476e-7)	369789.3962531291 iter/sec (stddev: 0.000001153966888084511)	0.64
benchmarks/test_guard_validate.py::test_g10_validate_deep_chain	30706.06571515804 iter/sec (stddev: 0.000027088054516283054)	24553.83231306249 iter/sec (stddev: 0.00006929897944825248)	0.80
benchmarks/test_guard_validate.py::test_g11_validate_wide	21185.489126277793 iter/sec (stddev: 0.000009578569179670694)	16452.432775735786 iter/sec (stddev: 0.000014109230742375827)	0.78

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