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 Galbaz1/video-research-mcp --skill hardware-evidence-capturegit 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/skills/galbaz1/video-research-mcp/hardware-evidence-capture)<a href="https://agentmods.dev/skills/galbaz1/video-research-mcp/hardware-evidence-capture"><img src="https://agentmods.dev/badge/skills/galbaz1/video-research-mcp/hardware-evidence-capture/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/galbaz1/video-research-mcp/hardware-evidence-capture"><img src="https://agentmods.dev/badge/skills/galbaz1/video-research-mcp/hardware-evidence-capture.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 | $0.00018 | $0.00339 |
| Opus 5.5 | $0.00007 | $0.00136 |
| Sonnet 5.5 | $0.00004 | $0.00068 |
| Haiku 4.5 | $0.00002 | $0.00034 |
Grade A, and why
hardware-evidence-capture 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.
What it actually says
Hardware evidence capture
Use mhs_discover and mhs_meta_info to inspect the named device and current
capabilities, directions and limits. All currently supported devices are
simulated. Preserve that label in every result; never infer physical state,
calibration, safety or camera coverage.
Read sensors or images with mhs_read and retain the returned original artifact
path, byte count, SHA256, command/state identity, method and units. Generated
simulator images are synthetic data and cannot become factual original evidence.
Do not upload or publish observations without separate authority.
For a write, provide one explicit command ID and exact parameters. Hard limits
cannot be overridden. confirm=true cannot approve a soft override or a
confirmation-required operation. Return the full approval proposal for an
operator's review; only the separately configured host file can authorize it.
Do not write that file on the operator's behalf without an explicit instruction.
An awaiting proposal has not attempted its action. A terminal command ID replays
its receipt and never re-executes.
After a failed write, inspect mhs_health_check before any new write. Inspect
the executed command, effective limits and resulting state in its receipt.
Emergency stop latches the simulator off. Soft recovery needs its own exact host
approval. Neither outcome proves physical stop or recovery effectiveness.
Physical adapters are unavailable pending named owner qualification. Stop on missing authority or an unknown result; preserve the error and receipt rather than inventing a device observation or automatically retrying.
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 · 34 lines · 18 tokens per session scan A e101c92f51e9
hardware-evidence-capture is a skill published in the GitHub repository Galbaz1/video-research-mcp (23 stars, last pushed 6d ago), licensed MIT. It adds 18 tokens to every session and 339 once invoked, about $0.0001 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 2026-10-03.
hardware-evidence-capture compared with skills from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| hardware-evidence-captureGalbaz1/video-research-mcp | Skill | 23 | 6d ago | MIT | A | 18 | Claude Code |
| meteocontrol-assistantnielsvbrecht/meteocontrol-mcp | Skill | 0 | 4mo ago | Apache-2.0 | A | 40 | Claude Code, Codex |
| network-health-checksirkirby/unifi-mcp | Skill | 1.0k | yesterday | MIT | A | 49 | Claude Code |
| video-perceptionjordanrendric/claude-video-vision | Skill | 1.5k | 23d ago | MIT | A | 51 | Claude Code |
| eg-precommit-reviewvshvedov/elephant-goldfish | Skill | 92 | 2d ago | MIT | A | 24 | Codex |
| b123d-modelingpzfreo/build123d-mcp | Skill | 142 | 22d ago | Apache-2.0 | A | 0 | Claude Code, Codex |
| farming-expertpersonamanagmentlayer/pcl | Skill | 44 | 11d ago | Apache-2.0 | A | 50 | Claude Code |