Part of #2560 — per-language tracking for the keyword-rosetta control corpus (built for #1096). Planted intent is identical in every language, so every deviation below is measured bias vs. the 46-language median, validated in the corpus's deviation ledger. 🔴 = beyond ±50% of the cross-language median, 🟡 = ±25–50%.
Triage order (per the epic): structure detection first, then signal/keyword rules, then risk-score consequences — §3's numbers are downstream of §1–§2 and re-baseline as those fix. Cross-cutting engine issues (e.g. #2535/#2546) fix many languages at once; this issue tracks this language's residual after each cross-cutting fix lands, and closes when the language sits inside the ±25% band on every comparable metric (or every remaining deviation is ledgered as intended morphology).
1. Structure detection (fix first)
In band — no structure-level deviations.
2. Signal / keyword deviations
- 🔴
safety: 4 vs median 2 (+100%)
3. Risk-exposure consequences (downstream)
In band — no risk-score deviations.
Ledgered shapes touching typescript
| defect |
type |
issue |
evidence |
summary |
census-requires-git-tracked |
engine-semantic |
— |
— |
GalaxyScope's census enumerates git-tracked files only; an untracked folder scans as '0 files mapped' with no per-file warning |
state-flux-branch-weighting |
engine-semantic |
#2546 |
b.ts |
state_mutation is recorded x3 when branch context exists in the file ('flux' semantics, matching the state_flux column name) |
ts-callparen-args |
engine-semantic |
— |
(notes) |
typescript's args rule counts statement-position call parens (describe(kit), expect(kit)) in addition to declaration params — a.ts args 5 for 3 declarations (finditer-co… |
string-literal-selective-shielding |
upstream-question |
#2535 |
b.ts |
Keywords inside one string literal ('if eval fails, try open') count toward branch (+2), safety (+1), io (+1) but eval does NOT count toward high_risk_execution (routes … |
Decoy outcomes (file-level evidence)
- b.ts ·
const note = "if eval fails, try open" (string): if + try count (branch +2, safety +1), eval does NOT (js/python selective profile); let+var 2 raw state tripled to 6. Ledger: string-literal-selective-shielding, state-flux-branch-weighting
- c.ts ·
const HACK_LEVEL = level (code): inert (underscore)
Reproduce / evidence
Part of #2560 — per-language tracking for the keyword-rosetta control corpus (built for #1096). Planted intent is identical in every language, so every deviation below is measured bias vs. the 46-language median, validated in the corpus's deviation ledger. 🔴 = beyond ±50% of the cross-language median, 🟡 = ±25–50%.
Triage order (per the epic): structure detection first, then signal/keyword rules, then risk-score consequences — §3's numbers are downstream of §1–§2 and re-baseline as those fix. Cross-cutting engine issues (e.g. #2535/#2546) fix many languages at once; this issue tracks this language's residual after each cross-cutting fix lands, and closes when the language sits inside the ±25% band on every comparable metric (or every remaining deviation is ledgered as intended morphology).
1. Structure detection (fix first)
In band — no structure-level deviations.
2. Signal / keyword deviations
safety: 4 vs median 2 (+100%)3. Risk-exposure consequences (downstream)
In band — no risk-score deviations.
Ledgered shapes touching typescript
census-requires-git-trackedstate-flux-branch-weightingts-callparen-argsstring-literal-selective-shieldingDecoy outcomes (file-level evidence)
const note = "if eval fails, try open"(string): if + try count (branch +2, safety +1), eval does NOT (js/python selective profile); let+var 2 raw state tripled to 6. Ledger: string-literal-selective-shielding, state-flux-branch-weightingconst HACK_LEVEL = level(code): inert (underscore)Reproduce / evidence
data/typescript/—expected_signals.jsoncarries the shape-by-shape accountingGALAXYSCOPE_BIN=<venv>/bin/galaxyscope python tools/verify_language.py typescriptin the corpus repo