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 bvisible/mcp-ssh-manager --skill ssh-restrictedgit clone --depth 1 https://github.com/bvisible/mcp-ssh-managerWrote 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/bvisible/mcp-ssh-manager/ssh-restricted)<a href="https://agentmods.dev/skills/bvisible/mcp-ssh-manager/ssh-restricted"><img src="https://agentmods.dev/badge/skills/bvisible/mcp-ssh-manager/ssh-restricted/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/bvisible/mcp-ssh-manager/ssh-restricted"><img src="https://agentmods.dev/badge/skills/bvisible/mcp-ssh-manager/ssh-restricted.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 | $0.00046 | $0.00551 |
| Opus 5.5 | $0.00018 | $0.00220 |
| Sonnet 5.5 | $0.00009 | $0.00110 |
| Haiku 4.5 | $0.00005 | $0.00055 |
Grade A, and why
ssh-restricted 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 3d 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 — 58 lines — stays where its author put it; the contents beside it link to each section on GitHub.
When a server says no
MCP SSH Manager lets an operator constrain what an agent may do, per server. A refusal is the system working, not a bug to route around.
The three answers you can get
"Policy denied: Tool X is disabled (mode: readonly)" — the server is
read-only. Mutating tools (ssh_deploy, ssh_upload, ssh_execute_sudo,
ssh_backup_restore, ssh_db_import…) are off entirely. Reading still works.
"Policy denied: ... matches built-in destructive pattern" or "does not
match the allow patterns" — the server is restricted. The operator listed
what may run there, and this is not on the list.
"Refused by the operator" — a human saw the command in the control plane and said no. That is a person, not a rule.
What to do about it
Say what was refused and why, plainly. The operator configured this on purpose; they need to know their agent hit the wall, not that "something went wrong".
Offer the read-only equivalent. Refused systemctl restart nginx? You can
still ssh_service_status, ssh_tail the error log, and report what you found
along with the exact command a human would need to run. That is usually more
useful than the restart would have been.
Do not look for a way around it. Do not:
- retry the same thing through
ssh_executebecause the specific tool was blocked — the policy layer covers both; - reach for another server that is less constrained to do the same work;
- ask the user to disable the mode so you can continue. If they want it off, they will say so.
Working around a control someone deliberately put in place is worse than failing the task, because they will believe it held.
Approval, and waiting
When a server is set to destructive or always approval, the engine pauses and
asks a human before running. From your side the call simply takes longer — that
is expected, and it is what the operator wanted.
If it comes back "Refused by the operator", do not run a variation of the same command hoping it slips through. Report the refusal and ask what they would prefer instead.
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.
- 3d ago First seen · 58 lines · 46 tokens per session scan A ae500c34f8ea
ssh-restricted is a skill published in the GitHub repository bvisible/mcp-ssh-manager (525 stars, last pushed 2d ago), licensed MIT. It adds 46 tokens to every session and 551 once invoked, about $0.0002 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 2026-10-26.
Other skills, from other repositories
re-port-flow
A procedure for filling Re:port Flow business templates and generating PDF documents. It supports invoices, quotations, and delivery notes, using only information supplied by the user.
aiss-consult
Consult prior AI war-stories before answering coding, deploy, debugging, refactoring, code-review, or framework-decision tasks. Pulls relevant context from the AI Success Story corpus via MCP — real incidents often catch gotchas parametric knowledge misses.
cortex-explore-memory
Explore the memory system's state, find gaps in knowledge, assess coverage, and get diagnostic information. Use when the user asks 'what does my memory look like', 'show me memory stats', 'what am I missing', 'how good is my knowledge', 'memory health', 'show coverage', 'find gaps', 'what topics are weak', or when you…
qmclaw-workbench
A local simulator for superconducting quantum-bit experiments, including measurements used to study and tune a qubit. It produces bounded synthetic results without connecting to laboratory hardware or quantum cloud services.
quantum-grover
A Qiskit implementation of Grover's algorithm, a quantum search method that increases the chance of finding a marked item. It can explain the method, build an explicit quantum circuit, or run it with chosen inputs.
quantum-algorithms
A routing workflow for choosing an open-source quantum-computing algorithm and loading the matching local workflow. Quantum computing uses circuits and quantum states to represent and process information.