Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add dcc-mcp/dcc-mcp-renderdoc --skill renderdoc-perfgit clone --depth 1 https://github.com/dcc-mcp/dcc-mcp-renderdocWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/dcc-mcp/dcc-mcp-renderdoc/renderdoc-perf)<a href="https://agentmods.dev/skills/dcc-mcp/dcc-mcp-renderdoc/renderdoc-perf"><img src="https://agentmods.dev/badge/skills/dcc-mcp/dcc-mcp-renderdoc/renderdoc-perf/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/dcc-mcp/dcc-mcp-renderdoc/renderdoc-perf"><img src="https://agentmods.dev/badge/skills/dcc-mcp/dcc-mcp-renderdoc/renderdoc-perf.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00119 | $0.01187 |
| Opus 5.5 | $0.00048 | $0.00475 |
| Sonnet 5.5 | $0.00024 | $0.00237 |
| Haiku 4.5 | $0.00012 | $0.00119 |
Grade A, and why
renderdoc-perf scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 1mo ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 91 lines — stays where its author put it; the contents beside it link to each section on GitHub.
RenderDoc Perf
These tools answer how expensive a frame is. They replay an existing .rdc through RenderDoc's
replay API and never modify it; only analyze_overdraw writes to disk, and only to a destination
you name.
Every tool in this skill needs the deep replay backend (renderdoc.pyd, reached through the
Python runtime bundled with qrenderdoc). The headless renderdoccmd baseline cannot read
anything back out of a capture, so it cannot substitute here.
Capability gating
Call perf_capabilities first. It reports:
- whether the deep backend is reachable, and how to enable it when it is not;
- when given
capture_file, the per-capture facts only a replay can answer — whether this driver exposes counters at all (counters), whether one of them carries GPU duration (timing), and whether post-VS geometry can be read back (post_vs_data); - per tool, whether that tool is usable right now and which flag it depends on.
When the deep backend is missing, every tool returns a structured
unsupported_backend result naming the missing backend and how to enable it. None of them raise,
and none of them return empty data silently.
| Tool | Needs | Answers |
|---|---|---|
list_counters |
— | which GPU counters this driver exposes, with unit and category |
fetch_counters |
counters |
sampled counter values over one event range |
get_action_timing |
timing |
per-action GPU duration, slowest draws, per-pass totals |
get_debug_messages |
— | debug, warning, and error messages a replay produced |
analyze_overdraw |
post_vs_data |
how many times the frame shades each pixel, with a heatmap |
Working order
perf_capabilities— confirm the backend and the per-capture flags.list_counters— see what this driver offers before asking for samples.fetch_counters— sample the counters you picked over the event range you care about.get_action_timing— turn the timing counter into per-action durations and per-pass totals.get_debug_messages— read the warnings and errors the API produced during replay.analyze_overdraw— check whether the frame is fill-rate bound.
What ships with it
8 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
- metadata/depends.md 18 B
- scripts/analyze_overdraw.py 1.7 KB runs code
- scripts/fetch_counters.py 1.5 KB runs code
- scripts/get_action_timing.py 2.0 KB runs code
- scripts/get_debug_messages.py 1.1 KB runs code
- scripts/list_counters.py 1010 B runs code
- scripts/perf_capabilities.py 617 B runs code
- tools.yaml 9.4 KB
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 1mo ago First seen · 91 lines · 119 tokens per session scan A d519f59c9328
renderdoc-perf is a skill published in the GitHub repository dcc-mcp/dcc-mcp-renderdoc (1 stars, last pushed yesterday), licensed MIT. It adds 119 tokens to every session and 1,187 once invoked, about $0.0005 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-21.
Other skills, from other repositories
renderdoc-mcp
Analyze RenderDoc GPU frame captures with renderdoc-mcp MCP tools. Use when Codex needs to inspect .rdc captures, diagnose black screens or visual artifacts, explain frame structure, inspect specific draw calls, or investigate GPU rendering and performance issues.
performance-optimization
Find and fix game performance problems methodically — measure with the engine profiler first, reason about the frame-time budget, locate the CPU-vs-GPU bottleneck, then apply the right fix: object pooling, draw-call batching, fewer allocations/GC spikes, and asset budgets. Engine- neutral method that pairs with each…
beehave
Use when using the Beehave addon — pure-GDScript behavior trees with composites, decorators, leaves, a blackboard, and a visual runtime debugger.
godot-debugging
Use when debugging Godot projects — remote debugger, print techniques, signal tracing, common error patterns and fixes.
rdc-cli
Use this skill when working with RenderDoc capture files (.rdc), analyzing GPU frames, tracing shaders, inspecting draw calls, or running CI assertions against GPU captures. Trigger phrases: "open capture", "rdc file", ".rdc", "renderdoc", "shader debug", "pixel trace", "draw calls", "GPU frame", "assert pixel"…
asset-audit
Audits game assets for compliance with naming conventions, file size budgets, format standards, and pipeline requirements. Identifies orphaned assets, missing references, and standard violations.