EPIC: T27 backend — rings from the decision core to silicon

[gHashTag]

EPIC: T27 backend - rings from the decision core to silicon

Everything below the interface is written in T27 and generated to its target. Swift keeps the interface and nothing under it. A rule transcribed into Swift, Rust, Zig and Verilog is four rules that agree until someone edits one.

Rings, innermost first

Ring	What	Depends	Target
T27-00	Decision core: retry, review, merge gate, capacity	-	Rust + Verilog
T27-01	A2A protocol	00	Rust + Postgres
T27-02	Orchestration: the Queen's tick	00, 01	Rust + inngest
T27-03	Transport: SSE	01	Rust
T27-04	Scoring: salience, reliability, latency	00	Rust + Verilog

Division of labour

Another agent owns /Users/playra/t27 and the FPGA boards. Do not edit /Users/playra/t27. Do not program a board. t27/fpga/HARDWARE_SSOT.md is authoritative for hardware.

Where it stands

RING-00 is executed on live data every round (#1331) and its Rust answer is compared to Swift's; the Swift answer is the one acted on. Five compiler gaps in the seed block generation of valid Rust for the other rings. Those gaps are the other agent's repository.

What a bee can do here without a Swift compiler, a board or the seed: write the parity ledger that says, per ring, what exists, what is generated, what is executed, and what decides - so the next wave starts from a fact and not from this table.

User Scenarios & Testing

User Story 1 - The ring table is read from the tree, not from this issue (P1)

Acceptance Scenarios:

1. Given the directory trios/rings/,
   When the ledger is generated,
   Then every T27-* directory present is listed with the count of .t27 files and of generated files under it.
2. Given RING-00's parity script trios/tests/t27/ring00_parity.sh,
   When the ledger is generated,
   Then the ledger states whether that script exists and how many assertions it contains, by counting.

Requirements

• FR-001: A script MUST walk trios/rings/ and emit one table row per T27-* directory, with file counts, not descriptions.
• FR-002: The ledger MUST record, for each ring, which of gen-rust, gen, gen-c, gen-verilog outputs exist as files, by name.
• FR-003: The ledger MUST NOT claim any ring "decides" anything; it records presence only.
• FR-004: The script MUST NOT read or write anything under /Users/playra/t27.

Success Criteria

• trios/tools/ring-ledger.sh exists, is POSIX sh, and produces trios/docs/ring-ledger.md.
• The ledger has one table with a header row and at least one row per T27-* directory found.
• The ledger contains the exact string ring00_parity.sh and a count of its test or grep -q lines.
• The script defines a function named ring_row; that identifier appears nowhere in the tree today.

Boundary

trios/tools/ring-ledger.sh
trios/docs/ring-ledger.md

[gHashTag]

Two of the five seed gaps are closed. gHashTag/t27#2247 — [GOLD-RING], branched from origin/master, seal moved per FROZEN.md §5, docs/NOW.md synced.

• Verdict.escalate now emits Verdict::escalate.
• The shorthand .escalate now takes its type from the function's declared return type instead of emitting escalate::.

Both forms compile: rustc --crate-type lib produces an rlib from each, and from the ring-00 source a bee wrote (14912 bytes). Before the change neither did.

Corrected finding. Gap 5 is not about comment style. pub module name; is rejected at the top level — "Unexpected top-level token: KwModule" — while module name; is accepted. test_highlight.t27 uses the pub form, which is why it reads as supported; that file is a highlighting fixture and is never compiled.

Still open in the seed: switch as a function body emits an empty match; ; comments inside an enum body are parsed as variants.

Swarm status. The Queen picked #1280 herself once she could see this epic, and put bees on #1280, #1281, #1282 and #1283. #1282's bee committed 288 lines of tests/t27/ring00_verilog.sh before timing out. Two record-keeping defects surfaced and are under investigation before anything is merged:

• RING-00: the decision core in T27, generating valid Rust #1280 is recorded accepted with committedFiles: 1, and its branch does not exist.
• RING-00: Verilog from the same source, checked by simulation #1282 is recorded failed with committedFiles: None, and its branch carries the 288-line commit.

Nothing from this epic will be merged until those two are explained.

[gHashTag]

Left OPEN deliberately during the 2026-09-04 backlog sweep. Seventy issues whose dispatch the Queen had accepted were closed and their branches pushed; this one was excluded because it is an EPIC and closing a parent while its children may still be open would hide unfinished work. Close it by hand once its children are accounted for.