docs(assets): record second shooter enemy tranche quality gate

[colorpulse6]

Summary

This is the quality-controlled second M2 shooter-enemy tranche for Cloaker, Scout, Wraith, and Echo.

All generated candidates were rejected by their visual or live-consumer gates, so zero production sprites are replaced. The PR records the audit, exact-size/live baseline evidence, complete generation provenance, rejection decisions, and the final eight-enemy production silhouette comparison.

Frequency audit

Counts are explicit authored campaign plus planet spawns; conditional boss additions are excluded.

Enemy	Campaign	Planet	Total
Cloaker	369	34	403
Scout	319	77	396
Wraith	296	44	340
Echo	296	44	340

These are the next four highest-frequency authored enemies after Swarm, Bomber, Gunner, and Drone.

Role-read and runtime contracts

Enemy	Intended role	Source/output	Frames	Alpha bounds	Entity	Draw	Bestiary	Representative level
Cloaker	Crescent/shutter stealth hunter, readable visible and at 15% alpha	1536x1024 sRGBA	1	633x596+452+147	44x44	52x52	96x96	W5-3 Wave 1
Scout	Lean, uncomplicated fast unarmed baseline interceptor	1536x1024 sRGBA	1	602x476+465+240	40x40	48x48	96x96	W1-1 Wave 1 and Wave 8
Wraith	Broad, heavy corrupted sarcophagus/reliquary	1024x1536 sRGBA	1	968x1284+28+97	48x44	56x52	96x96	W4-2 Wave 1
Echo	Compact connected displaced/repeating phase craft	1536x1024 sRGBA	1	1037x895+258+49	36x36	44x44	96x96	W6-2 Wave 1

Because every candidate was rejected, each output remains byte-identical to the original at the dimensions and bounds above.

Decisions

Enemy	Decision	Candidate reasoning
Cloaker	Original retained	Initial candidate could not fit the approved broad envelope without underfill/crop. The retry disappeared at 15% alpha in W5-3 and read right-facing while visible.
Scout	Original retained	Initial candidate collapsed to an undersized needle. The retry became nearly square and its twin prongs read as guns/heavy structure on an unarmed baseline attacker.
Wraith	Original retained	Initial candidate exposed a literal humanoid/skull. The retry fixed that defect and passed static/W4-2 checks, but live Bestiary detail could not be verified.
Echo	Original retained	Candidate passed exact-size, 15%-alpha, and W6-2 presentation, but live Bestiary detail could not be verified.

No misleading candidate after-* or matched acceptance sheets are included because no replacement was accepted.

Live verification

• W1-1 Wave 1: 5/5 Scouts readable in the sparse opening line.
• W1-1 Wave 8: 8/8 Scouts readable in the dense authored V.
• W5-3 Wave 1: 9/9 Cloakers observed hidden and visible; fire only in the visible state.
• W4-2 Wave 1: 4/4 Wraiths readable over dark and brighter blue-gray scene regions under Cinder tint; red fire observed.
• W6-2 Wave 1: 6/6 Echoes observed hidden and visible; fire only in the visible state.
• Live Bestiary detail verified for Scout and Cloaker.
• Wraith detail remained on the list despite a visible foreground page, active rAF sampling, and trusted Enter/Space delivery; Echo detail was not attempted after that confirmed blocker. On 2026-07-16 the user explicitly directed delivery to continue with this automation limitation documented. These checks remain unverified, not passed.
• Browser console: zero new errors. Two unchanged Next.js font-preload warnings appeared in the applicable session.
• All four production sprite requests returned HTTP 200 initially and 304 after reload, with no failures.

Verification

Run sequentially from game/:

• npx tsc --noEmit — pass
• yarn colony:test — 268/268
• yarn engine:test — 66/66
• yarn sprites:test — 4/4
• yarn build — static export pass
• NEXT_PUBLIC_DEVTOOLS=1 yarn build — DevPanel static export pass
• git diff --check origin/main...HEAD — clean
• Independent pre-PR review — approved, no findings

Restricted-sandbox IPC/font-fetch failures passed when the identical commands were rerun with the required local IPC/network permissions; no source or configuration changes were made.

Evidence and provenance

• Complete prompt/provenance ledger
• Final eight-enemy production silhouette comparison
• Exact-size baseline panels
• Representative gameplay baselines
• Bestiary baselines

Scope guarantee

No TypeScript, registrations, renderer defaults, atlas scripts, gameplay behavior, stats, spawning, visibility timing, collisions, hitboxes, balance, saves, bosses, first-person billboards, backgrounds, hubs, lockfiles, package metadata, or production sprite files changed.

Do not merge: visual/provenance approval pending.

[qodo-code-review]

Qodo reviews are paused for this user.

Troubleshooting steps vary by plan Learn more →

On a Teams plan?
Reviews resume once this user has a paid seat and their Git account is linked in Qodo.
Link Git account →

Using GitHub Enterprise Server, GitLab Self-Managed, or Bitbucket Data Center?
These require an Enterprise plan - Contact us
Contact us →