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.
git clone --depth 1 https://github.com/Galbaz1/video-research-mcpWrote 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/commands/galbaz1/video-research-mcp/traces)<a href="https://agentmods.dev/commands/galbaz1/video-research-mcp/traces"><img src="https://agentmods.dev/badge/commands/galbaz1/video-research-mcp/traces/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/commands/galbaz1/video-research-mcp/traces"><img src="https://agentmods.dev/badge/commands/galbaz1/video-research-mcp/traces.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 | $0.00013 | $0.00897 |
| Opus 5 | $0.00006 | $0.00449 |
| Sonnet 5 | $0.00003 | $0.00179 |
| Haiku 4.5 | $0.00001 | $0.00090 |
Grade A, and why
traces 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 9d 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 — 117 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Traces: $ARGUMENTS
Query and debug MLflow traces captured from video-research-mcp Gemini API calls.
Prerequisites
The MLflow tracking server must be running. If tools are unavailable or connection is refused:
MLFLOW_TRACKING_URI=http://127.0.0.1:5001 mlflow server --port 5001
Then restart Claude Code to reconnect the mlflow-mcp MCP server.
Default Experiment
Use experiment name video-research-mcp unless the user specifies otherwise. Resolve the experiment ID by searching with that name first.
Context Discipline (mandatory)
- Always use
extract_fields— never pull full traces. Gemini traces contain large payloads (video URIs, cached content, full prompts/responses). - Keep output compact: target <= 300 tokens for overview, expand on detail requests.
- Never print raw JSON blobs or full span attributes.
Argument Routing
No arguments (empty $ARGUMENTS)
Show a recent traces overview:
search_traces({
experiment_id: "<resolved>",
max_results: 10,
extract_fields: "info.trace_id,info.state,info.execution_duration_ms,info.request_time,info.request_preview"
})
Present as a compact table: trace ID (short), status, duration, time, and request preview.
errors
Filter for failed traces:
search_traces({
experiment_id: "<resolved>",
filter_string: "status='ERROR'",
max_results: 20,
extract_fields: "info.trace_id,info.state,info.execution_duration_ms,info.request_time,info.request_preview"
})
slow
Filter for traces over 5 seconds:
search_traces({
experiment_id: "<resolved>",
filter_string: "execution_time_ms > 5000",
max_results: 20,
extract_fields: "info.trace_id,info.execution_duration_ms,info.request_time,info.request_preview,data.spans.*.name"
})
Trace ID (starts with tr- or is a hex string)
Get trace detail:
get_trace({
trace_id: "<id>",
extract_fields: "info.*,data.spans.*.name,data.spans.*.status_code,data.spans.*.start_time_unix_nano,data.spans.*.end_time_unix_nano"
})
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.
- 9d ago First seen · 117 lines · 13 tokens per session scan A afeb6932b0c5
traces is a command published in the GitHub repository Galbaz1/video-research-mcp (23 stars, last pushed 1mo ago), licensed MIT. It adds 13 tokens to every session and 897 once invoked, about $0.0001 per session on Opus 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 2026-08-30.
Other commands, from other repositories
eg-fix-bug
Fix a bug using the elephant/goldfish workflow — problem doc, goldfish diagnosis check, failing test, fix, review, validate.
safe-decompile
Safely fingerprint and decompile an Android artifact through the sandboxed MCP server.
mosaic-agent-shader
Launch Ray — Mosaic Bridge shader expert (ShaderGraph, HLSL, node wiring, shader debugging).
latent-fix
Quick fix cho file đang mở hoặc file được chỉ định. Tối đa 1-2 patches.
setup-video-vision
Interactive setup wizard for claude-video-vision — configure backend, whisper, frames, and verify dependencies.
debug
Diagnose Ansible playbook issues from symptoms.