agent: builder model: opus max_turns: 70 budget_usd: 16 —
Score: a compiler defect found by y9 (#1027’s run); bugs before moving. Nothing in yaml waits for it (the y2 draft captures with text(..)), so it is small and first.
Facts the y9 builder paid for: a rustc error from a grammar (Mismatched types, Cannot find value) is the emitter’s or the checker’s: run nikaia check (passes here) and nikaia lower FILE (writes the .rs, shows the rule’s lowered type) first. The backend is winnow-grammar (pinned rev, cargo checkout, not changed here). nikaia test --both-settings --project tests/compiler runs over 10 minutes: run_in_background with timeout: 3600000. If other sessions’ uncommitted files block git pull --rebase, push fast-forward and report Pushed: rejected (the driver rebases).
Found by the yaml lane (y9, #1027’s report, 2026-10-31). Nothing in packages/yaml needs it; a compiler defect, fixed in the compiler (CLAUDE.md).
Probe (the architect ran each row below on 94f07e7b):
grammar G {
rule WS = "" -> { }
entry rule top -> String = x:"ab" eof -> { x }
}
fn main() {
let r = G::top("ab") catch {
println("error")
return
}
println(r)
}
nikaia check passes; nikaia run fails in rustc, Mismatched types at the action (over text and over a let-bound Bytes alike; the y9 builder saw Lifetime may not live long enough with a temporary Bytes(..) argument, which is #1004). The same with x:text("ab") prints ab. A choice of literals stored in a String field, x:("+" | "-") eof -> { T { s: x } }, fails the same way. Used inside the action only it works: n:"-"? ... -> { f"{n ?? "+"}{d}" } prints -12, +7 (and examples/1brc.nika:49).
CheckState::pattern_type (crates/nikaia-std/src/tools/check_grammar.nika:48) has no Pattern::Literal arm, so a literal’s label is Ty::Unknown; a rule -> String whose action returns a text(..) capture is lowered as -> &'a str (a view of the input, pattern_text()), one returning a literal capture keeps -> String and its &str value does not fit. The emitter already treats a literal’s capture as text (emit_pattern.nika:72, over bytes wrapped in text(..), ADR-366 D2). choice_type (same file) returns Unknown when its first alternative is, so the choice shape should follow from the same arm (unverified).
Part II 10.8’s table says "text" yields nothing; the row for x:p says a label binds what p yields. What the compiler does (a label on a literal binds the matched text) is the meaning in use (examples/1brc.nika); the table row gets that sentence.
crates/nikaia-std/src/tools/check_grammar.nika: CheckState::pattern_type (line 48), choice_type (line 75), pattern_text() (line 17).crates/nikaia-std/src/tools/emit_pattern.nika:72: the Pattern::Bind arm already wraps a literal in text(..) over bytes; read, change only if the test needs it.docs/specification/20-nikaia-advance.md:718: the "text" row of 10.8.crates/nikaia-std/src/tools/lowered/check_grammar.rs (and emit_pattern.rs), regenerated by lower-std (docs/builder-facts.md), never edited by hand.pattern_type: Pattern::Literal(_) => pattern_text(). Re-lower (lower-std, docs/builder-facts.md). Part II 10.8, the "text" row’s yields cell: nothing; a label on it binds the text matched. Test tests/compiler/tests/grammar_a_label_on_a_literal_binds_its_text.nika with .stdout: the probe over text, the same over a let-bound Bytes, the choice into a struct field, the optional one used in the action; it fails on the unfixed binary. Fixes this issue.Expected .stdout rows: ab (text), ab (bytes, input bound by let first: a temporary Bytes(..) argument is #1004, not this brief), - (choice), -12 and +7 (optional). The choice row is unverified as to its cause: if it still fails after the arm, stop, commit the rest without that row, and report it under Found:.
The new test passes and failed on the unfixed binary, the grammar_ tests pass, the gate of .claude/agents/builder.md (#852) is green, the commit says Fixes #1030. Changelog: one sentence (this changes crates/ and Part II); name the 10.8 sentence under Body: for the supervisor.
docs/builder-facts.md first. develop in the main checkout; git status at start: other agents’ uncommitted files are not yours, never touch or stage them.git add <files> only; commit after each step; git push origin develop fast-forward, no rebase (rejected: report Pushed: rejected, go on committing).cargo fmt, cargo clippy --all-targets -- -D warnings, scripts/pre-push-check.sh, and the gate of .claude/agents/builder.md (#852).CHANGELOG.md, badges; no new NK code. New tests in Nikaia only; never the full suite. No foreground sleep, no polling; scripts in files.std or the toolchain lacks stops the work: no code around it unless an ADR or Part I-III refuses it (name it). Else Result: blocked, Found: gap + probe, Needs: owner decision (CLAUDE.md, A missing form stops the work).NIKAIA_CACHE_LIMIT=2G; at most two clones with target/ in the machine; delete your own clone’s target/ and target/release/nikaia-scratch at the end; report GB freed.Pushed: line names the pushed sha. A report without a Pushed: line, or with a placeholder left in (TC, DISK), is not done and cuts the batch (k3a1 2026-10-29: Result: done, nothing committed).NIKAIA_CACHE_LIMIT=2G target/fast/nikaia test --project packages/yaml green, and cargo test --profile fast -p nikaia --test it -- fmt::every_nika_file (it formats every .nika under packages/). For a change only under packages/yaml/ this is the gate the line above asks for; a change under crates/ runs the gate of .claude/agents/builder.md (#852) as well. target/fast/nikaia is built by the session hook; if it is missing, cargo build --profile fast -p nikaia once.Changelog: none, unless the brief changes crates/.Files you may change: crates/nikaia-std/src/tools/check_grammar.nika (and emit_pattern.nika if the test needs it), their lowered twins and the index lowered.rs, the "text" row of Part II 10.8 (docs/specification/20-nikaia-advance.md), the new test and its .stdout, your report. Nothing else; never docs/briefs/batch.txt or a brief of another lane.
max_turns: commit and push every finished step, then report Result: partial with what is left under Needs:. Uncommitted work at the end is lost work (y4 ran out at 110)./home/user/Nikaia/docs/briefs/yaml/y10-1030-literal-label.md.report.md)Fields of docs/architect.md, Briefs and reports: Result: done|noop|partial|blocked (the bare word alone on that line: Result: done, never a summary after it, which cuts the batch, b869a 2026-10-27; Owner: is a question for the owner or none, never a role), Files, Commit, Body (optional, one paragraph for the commit message), Fixes, Pushed, Tests, Gate (green|red), Deviation, Found, Needs, Changelog, Owner, Disk freed, and for a move Rs2nika: (the kept N of M lines of rs2nika --score, then the three costliest hand edits; none for a task that moves no Rust, none (redesign) for a brief with Shape: redesign and no Convert: line). none where nothing, and then the bare word: Deviation: none (a note) cuts the batch (batch.nika compares exactly); a note goes into the commit message. Found: is only for a defect outside your change (it cuts the batch); a process event such as another session resetting or rewriting your files is no Found:, it goes under Needs: (m3a 2026-10-28); a converter draft that came out wrong is no Found:, it is one of the costliest edits under Rs2nika:, which the converter lane reads (e12-4b 2026-10-27); how you solved your own step goes into the commit message. Deviation: is only a step of the brief built otherwise than written or not built (a choice the brief left to you, such as keep what X needs and say so, is no deviation: it goes into the commit message; nor is building two steps in one commit when every step is built, e12-3a 2026-10-27, nor a step another commit on develop already built, shown by your test failing with that commit’s change reverted, b889 2026-10-27: name the commit in Body:, nor a test’s expected .stdout/.stderr taken from what the program really prints where the brief’s sketch of it was short or wrong, b934 2026-10-29: say so in Body:, nor a lowering that follows the ADR governing it where the brief’s expected lowering contradicts that ADR, b933 2026-10-29: name the ADR in Body:, nor re-keying an allow-list entry (KNOWN) whose hash changed for the same program and the same reason, b-fixforward-sweep2 2026-10-28: say so in Body:); how you verified (e.g. a test not run against the unfixed compiler) goes under Tests:. A noop says why under Tests: (what was probed), not under Deviation:. Final answer, one line: done <sha>, report <path> (or BLOCKED: ... / PUSH REJECTED ...).
Attribution: Co-Authored-By: Claude Opus 5.5 noreply@anthropic.com Claude-Session: https://claude.ai/code/session_01UVxJDveRM5qKZKyCqq6Y24