docs: release notes and version scoping for v1.3.0 - #348

What

The engine's 1.3.0 release notes and version scoping, docs only (CHANGELOG.md, README.md, LEGAL.md,
docs/pages/index.md, docs/pages/install.md, docs/pages/public-contract.md, docs/strategy-capabilities.md,
runner/README.md). The release workflow writes only VERSION; the notes and scoping come from this change.

• CHANGELOG ## 1.3.0 — 2026-10-06. A minor release paired with pineforge-codegen 1.3.0 (hub image
  ghcr.io/pineforge-4pass/pineforge-release:1.3.0, the three pineforge-v1.3.0-* tarballs with .sha256).
  pineforge-live admits the proven confirmed-bar request shapes and standalone close-only varip from the
  confirmed-bar receipt a codegen 1.3.0 strategy exports (Live runner: admit proven confirmed-bar request shapes and standalone varip #342); a codegen 1.2.0 module keeps 1.2.0's refusals,
  and process_orders_on_close stays refused. Pine adapter parity changes (Pine adapter: limit reach, exit binding, carried trail, margin-call schedule #341, Pine adapter: margin calls at the open, the exact reversal tie, and global exits on a held position, pinned by TradingView controls #345, Pine adapter: the same-pass position view after a close fill, the close's owed follow-up, open-marketable exits first, and same-side global exits, pinned by TradingView controls #347), stating plainly
  the global exit (no from_entry) that now binds to the held position beside an opposite entry order (Pine adapter: margin calls at the open, the exact reversal tie, and global exits on a held position, pinned by TradingView controls #345) and
  beside a same-side add in pyramiding scripts (Pine adapter: the same-pass position view after a close fill, the close's owed follow-up, open-marketable exits first, and same-side global exits, pinned by TradingView controls #347), and the margin calls at the open after a
  process_orders_on_close close fill (Pine adapter: margin calls at the open, the exact reversal tie, and global exits on a held position, pinned by TradingView controls #345). Tests and docs (docs: render the active scoreboard from the facts tokens at pineforge-release 3e39d64 #343, Exit binding tests: pin that the cap part judges an add once #344) in one line.
• Version scoping. README (install line, prebuilt v1.3.0 tarballs, the pairing sentence, codegen releases
  1.0.0 to 1.3.0, the runtime paragraph, the benchmark line now names v1.3.0 as not re-measured, a v1.3.0
  Releases line), install.md (v1.3.0 tarballs, engine1.3.0-codegen1.3.0), public-contract.md (the 1.3.0
  release sentence, the 1.3.0 pair, C-SURFACE-2 "none of 1.1.0, 1.2.0 and 1.3.0 includes them"). Live runner: admit proven confirmed-bar request shapes and standalone varip #342's
  "unreleased; after 1.2.0" markers become "since 1.3.0" (README, strategy-capabilities.md, runner/README.md).
  Lines that state what an earlier release shipped are kept. The release-scoreboard sentence and every
  <!-- pf: --> marker are untouched.
• Codegen license lines (README, LEGAL.md, docs/pages/index.md): the PineForge Source License 1.2 from
  codegen 1.3.0; release 1.2.0 keeps the PineForge Source License 1.1, releases up to 1.1.0 the PolyForm
  Noncommercial terms they shipped with. README and LEGAL.md add, in the LICENSE's terms, that the capital in an
  account a proprietary-trading firm or funded-trader program provides or allocates, including a challenge,
  evaluation or simulated account, is investment capital.

Report keys

No report key changes. docker/run_json.py is byte-identical v1.2.0 → main, so the harness the hub image runs
adds, removes and renames no key; EQUITY_REPORT_KEYS is unchanged; pf_report_t and pf_trade_t are unchanged
(<pineforge/pineforge.h> is byte-identical). The runner's versioned formats are unchanged; pineforge-live run's
stop summary adds a confirmed_bar_capabilities member (null without the receipt).

State hashes

Every recipe change since v1.2.0 is an addition in src/source/pine_state_hash.cpp (+46/−0), each folded only
while its new state is set; nothing removed, renamed or re-encoded; pineforge-broker-state/v19,
pineforge-source-adapter/v4, the stream fingerprint's version 19 and the epoch engine_script_run_v19 are
unchanged.

• Pine adapter: limit reach, exit binding, carried trail, margin-call schedule #341: the pending_bound_exit placement flag; the id-bound exit calls kept after a strategy.cancel; a short's
  owed lagged follow-up; a margin call booked at a close after the script; the carried trails re-priced for the
  bar in flight.
• Pine adapter: the same-pass position view after a close fill, the close's owed follow-up, open-marketable exits first, and same-side global exits, pinned by TradingView controls #347: the opening margin checkpoint owed after the exit legs marketable at an open.
• Pine adapter: margin calls at the open, the exact reversal tie, and global exits on a held position, pinned by TradingView controls #345: no recipe change (two transient members, waived).

check_aggregate_cpp_versions.py: "aggregate v19 ownership verified" at v1.2.0 and main.
check_broker_state_hash_coverage.py: OK at both (main: 277 source-adapter members, 279 nested fields, 79 waivers;
v1.2.0: 262, 278, 76).

Verified

• Tutorial numbers re-measured on an AWS spot box at this branch's code (engine 7a1f01c from a git bundle, GCC
  13.3; codegen main d1220bb for the README path): bash tutorial/run.sh 50 trades (17W / 33L, 34.0% win),
  +569.97, best/worst +1149.00 / −1111.97, max dd −4045.15; run_advanced.py and the Python sweep page's 12 rows
  as documented; run_stream.py 50 trades, +569.97; the README's own commands (install, transpile with codegen
  main, compile, run_strategy.py) 50 trades, net_profit=569.97. Unchanged, so no number is edited. The box is
  terminated.
• check_doc_reverts.py --base origin/main OK, check_doc_anchors.py (1,744 resolve), check_doc_lint.py
  clean, git diff --check clean; no changed path is in ci_docs_only.py's PROOF_READS.
• The merge gate's docs basis (preflight + docs) passed on the executor's commit 0e1ced7: ci_preflight: passed (exit 0, 44 stages), Doxygen warnings: 0; it is re-run on this PR's head for the pineforge/verify status.
• An independent read-only fact-check of the CHANGELOG section against the squash bodies and the tree; its seven
  scope corrections are applied.

Release grades (second commit)

The README's release sentence moves to the releases[1.3.0] tokens, its main-scoreboard lines (README, docs/pages/contributing-llm.md) re-render, and the 1.3.0 changelog says which registry baseline grades the release, all rendered from pineforge-release facts/facts.json at ff21de0 (#32): release 1.3.0 is graded on pineforge-parity-baseline-20261006-engine-7a1f01c0 (engine 7a1f01c, the merge of #347, which v1.3.0 equals in behaviour), 7,982 excellent / 7 strong of 7,989 probes, 0 below strong. lab facts check and --github-render pass; check_doc_reverts.py names the commits whose lines it re-renders.

🤖 Generated with Claude Code