AgentStack
Browse Sign in
Browse Why AgentStack Sell Docs
Sign in
SKILL verified Zlib Self-run

Dev Explain

skill-nebulavenus-forge-gpu-dev-explain · by Nebulavenus

Explain something in detail with investigation. No changes, no suggestions, no opinions. Facts and technical depth only. Usage: /dev-explain <topic>

No reviews yet
0 installs
5 views
0.0% view→install

Install

$ agentstack add skill-nebulavenus-forge-gpu-dev-explain

✓ scanned · ✓ verified, works with Claude Code, Cursor, and more.

Security review

✓ Passed

No issues found. Passed automated security review. · v0.1.0 How review works →

  • Prompt-injection patterns
  • Secret / credential exfiltration
  • Dangerous shell & filesystem operations
  • Untrusted network calls
  • Known-malicious package signatures

What it can access

  • Network access No
  • Filesystem access No
  • Shell / process execution No
  • Environment & secrets No
  • Dynamic code execution No

From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.

View the full security report →

Verified badge

Passed review? Show it. Paste this badge into your README, it links to the public security report.

AgentStack Verified badge Links to your public security report.
[![AgentStack Verified](https://agentstack.voostack.com/badges/verified.svg)](https://agentstack.voostack.com/security/report/skill-nebulavenus-forge-gpu-dev-explain)

Reliability & compatibility

Security review passed
0 installs to date
no reviews yet
4mo ago

Declared compatibility

Claude CodeClaude Desktop

Compatibility is declared by the source manifest. End-to-end runtime verification is coming, see below.

Preview Execution monitoring

We're building live execution health for every listing: tool-call success rate, median latency, uptime, and last-checked timestamps, measured, not self-reported. It isn't live yet, so we don't show numbers we can't stand behind.

How agent discovery & health will work →
Are you the author of Dev Explain? Claim this listing to set pricing, connect Stripe payouts, and keep 70% of every sale.
Sign up to claim

About

The user's topic is in the argument string passed to this skill. Investigate and explain it.

If no argument was provided, ask the user what they want explained.

Rules

  1. Explain with depth. The user wants to understand how something works.

Trace the full path: which functions are called, what data flows where, what the format looks like on disk, what the GPU receives. Include line numbers, struct layouts, byte offsets — the level of detail a programmer needs to modify the code confidently.

  1. Investigate. Unlike /dev-ask, you should read files, follow call

chains, check implementations, and verify claims against the actual code. The user expects you to do the legwork and report what you found.

  1. Do not make changes. Do not edit files. Do not propose edits. Do not

create branches. Do not commit. Read-only.

  1. Do not suggest. No "you might want to," no "consider," no "it would

be better to." The user asked how something works, not what to do about it.

  1. Do not apologize. There is nothing to apologize for.
  1. Do not be defensive. If something is broken, poorly designed, or

missing, say so plainly. The user needs accurate information, not reassurance.

  1. Report what is, not what should be. Describe the current state of the

code. Do not editorialize about whether it is good or bad. Do not propose alternatives unless the user explicitly asks "what are the alternatives."

  1. No personality. No greetings, no sign-offs, no emojis, no enthusiasm.

Write like a technical reference document.

Format

Start with a one-sentence summary of the answer. Then provide the detailed explanation with code references, data flow, and supporting evidence. Use headers to organize sections if the explanation covers multiple components. Include file paths and line numbers for every claim about the code.

Example invocation

/dev-explain how the scene renderer handles transparent submeshes

Response:

forge_scene_draw_model() in common/scene/forge_scene.h splits submeshes into two passes by alpha mode.

Pass 1 — Opaque and MASK

Iterates all nodes. For each node with a mesh, iterates submeshes. If material.alpha_mode is OPAQUE or MASK, draws immediately with SDL_DrawGPUIndexedPrimitives. MASK materials bind the shadow_mask pipeline which samples the base color texture and calls clip(alpha - cutoff) in the fragment shader (common/scene/shaders/shadow_mask.frag.hlsl:18). Depth writes are enabled.

Pass 2 — BLEND

Submeshes with alpha_mode == BLEND are collected into a ForgeSceneTransparentDraw array (max 256 entries). Each entry stores the node index, submesh index, and a sort distance computed from the submesh's precomputed object-space centroid transformed to world space via mat4_multiply_vec4(placement * node->world_transform, centroid). The distance is vec3_length(world_centroid - cam_position).

The array is sorted back-to-front with SDL_qsort using a float comparison. Then each entry is drawn with the blend pipeline: SRC_ALPHA / ONE_MINUS_SRC_ALPHA, depth test enabled, depth writes disabled.

Centroids are precomputed at model load time in forge_scene_load_model() by averaging vertex positions per submesh.

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

Install and usage instructions live in the source repository linked above.

Reviews

No reviews yet, be the first.

Versions

  • v0.1.0 Imported from the upstream source.