fusion
This server does not currently have a deployable release on Glama. Follow the repository's setup instructions to install and configure it for your MCP client, including any required credentials and local dependencies.
Start here
Read the case study: the problem, the headline result, a negative result, the limits.
Run the command below: it re-derives two committed runs offline, with no model and no key.
Check the Trust Report: every question asked of Fusion itself, generated from
evals/. The ledger shows each registration was committed before its evidence.Design: architecture.
Related MCP server: agent-firewall
Verify a real run
pip install -e .
fusion run verify --workspace examples/verified-run latest
fusion run verify --workspace examples/verified-run-toolloop latestverified: result re-derived exactly (in-sample guard verdicts excluded)Re-derives two real runs offline from their committed recordings: llama3.2:1b on the text corpus and qwen2.5:7b
through the tool loop (--corpus v2), both graded by Claude Code host subagents; exits non-zero on drift. What each
run shows: examples.
Use
pip install "fusion-safety[serve,mcp]"
fusion doctor
fusion run start --prompt agent.txt --target ollama:llama3.2:1b --grader claude-cli
fusion run finalize --min-grade B # report card; exit 1 below the bar
fusion serveAny model or agent can be the target: ollama:<model>, openai-compat:<url>#<model>, claude-cli:<model>, a
hosted API with your own key (openai:, anthropic:, hf:, ...; needs FUSION_ALLOW_API_SPEND=1), or any
program via cmd:<command>.
Test a tool-using agent (excessive_agency and data_exfiltration only):
fusion run start --prompt agent.txt --target ollama:qwen2.5:7b --grader host --corpus v2 --check excessive_agency,data_exfiltration
fusion run start --prompt agent.txt --target ollama:qwen2.5:7b --grader host --corpus v2 --check excessive_agency,data_exfiltration --no-fix-armEach attack offers your agent the probe's own tools and runs a bounded loop (at most four assistant turns); scripted
tool results come back, the injected one at its scripted turn. The real tool calls are graded by the oracle
cross-check and the judge, and the guard is replayed over every tool result the agent received. --corpus v2 also
adds gold v2's 80 sealed known-answer questions to a host-graded run; --no-fix-arm halves the target calls and
reports the as-written rate with its interval. Details: Tool loop.
Runtime guardrail in your agent loop:
from fusion_first.guardrail.guard import Guardrail
from fusion_first.guardrail.policy import GuardConfig
guard = Guardrail(GuardConfig(allowlisted_domains=["your-co.com"], require_authorization=True))
outcome = guard.guard_tool_call(tool_name, tool_args, user_request=user_message,
untrusted_context=True, untrusted_text=tool_results)
if outcome.blocked: ...Optional, off by default: Guardrail(config, content_safety=LlamaGuard()) adds a content-safety check on the
user's prompt (guard.guard_prompt(prompt); Llama Guard 3 on local Ollama, from fusion_first.guardrail.stages).
Optional, off by default: Guardrail(config, binder=RequestBinder("qwen2.5:7b", explain=True)) asks a local
Ollama model whether the user's request covers a tool call the rules would block as not covered. It sees the
request and the call, never the tool results, and every other rule still runs (fusion_first.guardrail.binder).
Request binding, with or without the binder, does not stop a hijacked call to the tool the user asked for with
other arguments (Trust Report, the informed-attack study): pair it with a confirmation step for consequential tools.
Claude Code: claude plugin marketplace add "$(fusion plugin-dir)" && claude plugin install fusion@fusion-first.
MCP: { "mcpServers": { "fusion": { "command": "fusion-mcp" } } } exposes the run engine and the guard's
guardrail_snippet / check_tool_call tools.
Registered studies
The tool-result injection detector study is pre-registered; its result is on the Trust page (registered result: not adopted; the detector stays dev-only).
The guard's optional content-safety stage is pre-registered; its result is in the Trust Report (registered result: with the stage the guard flags significantly more unsafe prompts, at a cost in wrong flags on safe prompts written to look unsafe; a category wording tuned on seen prompts was not adopted).
The judge's accuracy on tool trajectories is pre-registered; its result is on the Trust page (registered result: it met its registered bar). The version-three wording (claimed actions, pasted secrets) is registered and scored the same way.
License
MIT
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
This server cannot be deployed
Maintenance
Related MCP Connectors
Run AI evaluations, record agent traces and gate releases on evidence, with scoped org access.
Runtime permission, approval, and audit layer for AI agent tool execution.
Run AI agent tasks in isolated sandboxes with a cost cap, saved workflows and scoped connections.
Deterministic runtime safety for AI agents: scan PII, gate tool actions, verify LLM output.
Related MCP Servers
- AlicenseNot gradedqualityBmaintenanceEnables governed tool-calling agents with policy decisions, optional human approval, hash-chained audit logging, and deterministic evaluation.MIT
- AlicenseAqualityBmaintenanceEnables AI agent tool calls to be intercepted and judged before execution, allowing, asking a human, or blocking based on calibrated risk and policy.15MIT
- AlicenseNot gradedqualityNot gradedmaintenanceEnables monitoring and interruption of autonomous agent runs by detecting infinite tool loops, livelocks, and planning deadlocks, with safety guardrails for prompt injection, spend limits, and sandboxed command execution.MIT
- AlicenseAqualityCmaintenanceEnables AI agents to run shell commands, perform sandboxed file operations, and make HTTP requests through zero-trust guards that block command injection, path traversal, and SSRF while redacting secrets from inputs and outputs. Circuit breakers, per-session rate limits, and an append-only audit ledger prevent runaway loops, excessive token spend, and provide post-hoc traceability.5MIT