docs: rewrite for voice consistency and factual accuracy

[VijitSingh97]

A voice + accuracy pass over every prose doc, matching the cadence of the p2pool / xmrig / monero READMEs: one factual opening line, imperative procedures, plain NOTE: warnings, concrete numbers/ports/flags, link out instead of inlining, zero marketing.

Primarily voice. Technical facts, commands, flags, addresses, and ports are unchanged except where a doc was factually wrong versus the code — those are corrected and listed below. Every command, flag, default, and port was verified against the actual source (pithead, docker-compose.yml, config.reference.json, config.py, Dockerfiles, Makefile, test scripts). One commit per file for independent review.

Not touched (already in voice, or out of scope): docs/benchmarks/tor-vs-clearnet.md, tests/integration/README.md, tests/integration/testbench-README.md (no changes needed); CHANGELOG.md + docs/test-inventory.md + generated/vendor/template files (excluded by design).

1. Files changed (21)

File	What shifted
README.md	Flat spec-driven intro; tightened quickstart. No commands/facts changed.
docs/README.md	One-line factual pointer opening; cut the "new here?" softener.
docs/getting-started.md	Factual opening line; body already imperative.
docs/configuration.md	Tightened intro/prose; the config-key reference table left as-is (already code-accurate).
docs/workers.md	Dropped marketing cadence; tables/commands untouched.
docs/architecture.md	Flat opening; removed pitch framing; mermaid diagram + facts intact.
docs/privacy.md	Tightened narrative; fixed one stale checklist item (below).
docs/operations.md	Factual lead; every pithead subcommand re-verified against the script.
docs/dashboard.md	Factual voice; fixed the Syncing... badge text (below).
docs/hardware.md	Spec-driven prose; sizing verified against two live boxes; fixed the Tari overstatement + an mdb_copy error.
docs/testing-strategy.md	Factual cadence; removed a stale hard test-count.
docs/testing-guide.md	Fixed a wrong test path; retitled "Philosophy" → "Principles".
docs/integration-testing.md	Light tightening; all run.sh flags verified.
docs/releasing.md	Spec prose; corrected VERSION, image names, Tari pin, stale claim (below).
docs/release-server.md	Tightened opener; commands/paths verified.
docs/test-server-architecture.md	Voice; corrected the LMDB compaction command (below).
docs/faq.md	Removed "vs. doing it yourself / vs. Gupax" pitch framing; facts verified.
CONTRIBUTING.md	Imperative voice; corrected the branch model main → develop (below).
SECURITY.md	Removed pitch framing; reporting process kept verbatim.
build/dashboard/README.md	Corrected frontend-test claims; added sim/; fixed the node --test invocation (below).
images/launch/README.md	Light touch; dropped the "Marketing/" label.

2. Accuracy corrections

File:line	Claim as written	What the code says	Source
docs/privacy.md	"Route P2Pool outbound through Tor by editing its compose command:"	P2Pool is Tor-by-default in v1.1; pithead injects --socks5 <prefix>.25:9050 --socks5-proxy-type tor. Knob is p2pool.clearnet (default false). No compose editing.	pithead:1149, config.py:132
docs/dashboard.md:15	Syncing… (Unicode ellipsis)	Code emits Syncing... (three ASCII dots)	web/views.py:682
docs/testing-strategy.md:22	"The dashboard has ~140 unit tests"	~600 dashboard test functions today; stale figure removed (no hard count to rot)	grep 'def test_'
docs/testing-guide.md:38	tests/service/test_data_service.py	Actual path build/dashboard/tests/service/test_data_service.py	repo layout
docs/releasing.md	"VERSION is 0.1.0"	VERSION is 1.0.3	VERSION:1
docs/releasing.md	"one remaining piece is wiring ${STACK_VERSION} into compose"	Already wired	docker-compose.yml:20,51,…
docs/releasing.md	Image names dashboard, p2pool, …	Published as pithead-*	docker-compose.yml:20,51,…
docs/releasing.md	Tari minotari_node:vX-mainnet	Pinned v5.3.1-mainnet by digest	docker-compose.yml:107
docs/test-server-architecture.md	Compaction via mdb_copy -c	monero-blockchain-prune --copy-pruned-database (mdb_copy fails MDB_VERSION_MISMATCH on Monero's patched LMDB)	tests/integration/compact-chain.sh:15-16,49
CONTRIBUTING.md:32,69	Branch off / target main	Integration branch is develop (main is released-only)	.github/workflows/ci.yml:5, Makefile:12
build/dashboard/README.md:30-33	"Component rendering has no unit tests; manual browser smoke test"	tests/frontend/components.test.mjs renders every card through App against a real build_state() fixture	build/dashboard/tests/frontend/
build/dashboard/README.md:48	Layout tree omits sim/	mining_dashboard/sim/donation_model.py exists	dir listing
build/dashboard/README.md:68	node --test tests/frontend/*.test.mjs	CI uses node --test tests/frontend/	test headers
docs/hardware.md	"Tari's chain rivals a full Monero node"	Tari ~140 GB vs a full Monero node ~265 GB — it only rivals/exceeds a pruned node (~100 GB)	live: pithead-prod 139 GB, gouda 138 GB / 266 GB
docs/hardware.md	Reclaim space on a full chain by compacting with mdb_copy -c	Monero's patched LMDB rejects mdb_copy; use monero-blockchain-prune	tests/integration/compact-chain.sh:15-16
docs/hardware.md	Tari disk ~135 GB	measured ~138–139 GB and growing → ~140 GB	live: pithead-prod/gouda

3. Open items

• docs/hardware.md — the sizing figures were verified against two live 31 GB-RAM deployments (pithead-prod pruned, gouda full): Monero ~102 GB pruned / ~266 GB full and Tari ~139 GB confirm the documented disk sizes; the per-service RAM figures are provisioning minimums (real RSS is far lower). The [TODO] is resolved and removed.

Do not merge — for review.