rosetta[tcl]: cross-language consistency — 2 red / 3 amber metrics, 4 ledgered shapes

[squid-protocol]

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

• 🔴 io: 8 vs median 3 (+167%)
• 🟡 safety: 3 vs median 2 (+50%)
• 🟡 safety_bypasses: 3 vs median 2 (+50%)
• 🟡 state_mutation: 9 vs median 6 (+50%)

3. Risk-exposure consequences (downstream)

• 🔴 risk_cognitive_load: 5.38304 vs median 15.1444 (-64%)

Ledgered shapes touching tcl

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.tcl	state_mutation is recorded x3 when branch context exists in the file ('flux' semantics, matching the state_flux column name)
batch4-dual-keyword-overlaps	keyword-overlap	—	—	Collective entry for batch 4's finditer-confirmed dual-rule keywords, each detailed in its language's manifest notes: agc RESUME (cleanup+branch), livecode global (globa…
string-literal-selective-shielding	upstream-question	#2535	b.tcl	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.tcl · set note "if eval fails, try open" (string): if + try count branch (tcl branch includes try; try also safety +1), open counts io +1, AND eval counts safety_bypasses +1 (tcl eval-as-bypass counted from literals, like shell/ruby/matlab); the set itself is a 3rd raw state hit, tripled to 9. Ledger: string-literal-selective-shielding, state-flux-branch-weighting

Reproduce / evidence

• Corpus folder: data/tcl/ — expected_signals.json carries the shape-by-shape accounting
• Gate: GALAXYSCOPE_BIN=<venv>/bin/galaxyscope python tools/verify_language.py tcl in the corpus repo
• Full context: findings_by_language.md#tcl · bias_report.md · bias chart