Cinch MCP Server
Cinch MCP Server lets an MCP-compatible agent query AI security controls and assess agent deployments.
List available security checklists with item counts and frameworks.
Run a named checklist against an AI agent or deployment description.
Retrieve a specific checklist control by checklist name and item ID.
Fetch step-by-step security protocols.
Look up controls mapped to frameworks such as NIST AI RMF, OWASP, ATLAS, CUSTODY, and LASM.
Search all checklists for controls that mitigate a given threat.
Compare two checklists to find coverage gaps and duplicate controls.
Probe a running deployment's host, project, or endpoint to collect an evidence bundle.
Grade an evidence bundle to get a score, letter grade, insights, recommendations, and an action plan.
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@Cinch MCP Server run the agent containment checklist for my AI agent deployment"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
๐ก๏ธ Cinch
Contain your AI agents before they reach production.
MCP server + cross-harness skills for building and operating AI agents safely.
The model proposes; the architecture authorizes and enforces.
121 | 6 | 9 | 6 | 5 | 4 | 3 |
controls | checklists | MCP tools | crosswalks | protocols | templates | skills |
Grounded in NIST AI RMF ยท CISA Secure AI Development ยท OWASP LLM & Agentic ยท OWASP Agentic AI Top 10 ยท MITRE ATLAS ยท CUSTODY ยท LASM ยท Google SAIF ยท DORA 2025
AI agents can read data, invoke tools, execute code, call APIs, and initiate business processes. When they go wrong โ through prompt injection, excessive autonomy, credential theft, or model error โ the consequences are only as severe as the environment allows. Cinch gives you the enforceable controls, protocols, and runtime queries to keep that environment contained.
โจ What's inside
What | Why it matters | |
๐ | 6 checklists (121 controls) | Enforceable, framework-mapped YAML โ not advice, controls |
๐ | Probes (host ยท project ยท behaviour) | Assess a running agent: inspect its process, its deployment, and how it behaves under attack |
๐ก | MCP server (9 tools) | Any MCP-compatible agent queries controls at runtime โ and can collect + verify evidence |
๐ฅ๏ธ | Assessment console | Grade a deployment, get insights + a phased action plan, export JSON/CSV/PDF |
๐ง | 3 skills | Drop-in Hermes skill definitions for security reviews & red-teaming |
๐ | 5 protocols | Step-by-step deployment, incident-response, and evidence collection |
๐ | 6 mappings | Crosswalks: NIST AI RMF ยท OWASP LLM ยท OWASP Agentic ยท MITRE ATLAS ยท CUSTODY ยท LASM |
๐ | 4 templates | Security review, risk assessment, agent policy, deployment ADR |
Related MCP server: pyntra
๐ฅ๏ธ The assessment console
โถ Try the hosted console โ no install required.
One page, two modes: a landing page for the project, and a live containment assessment dashboard. It is not a static mockup โ it loads a real assessment result pack and renders the analysis: letter grade, insights, ranked recommendations, and a phased action plan. Mark a control Enforced / Gap / N-A and every one of those re-derives instantly.
View | What you get |
Overview | Grade, containment score, top insights, "do now" queue, open-gap feed, per-checklist counts |
Insights & plan | Grade + why it was capped, ranked findings, one recommendation per gap (do this โ verify like this), remediation phased Now / Next / Later with exit criteria |
Assessment | Walk the checklist control by control, attach evidence as you go |
Control library | All 121 controls across 6 checklists โ searchable, filterable, assessable inline |
Threat model | LASM 7ร4 attack-surface grid derived live from the catalog's layer mapping |
Framework coverage | NIST AI RMF, OWASP, CUSTODY, ATLAS, LASM crosswalks against your answers |
Export | JSON (full result pack for CI gating and diffs), CSV (gap register with phase + rank), PDF (grade, insights, action plan, recommendations, crosswalk) |
Where the results come from
Two ways to produce them, one engine grading both:
Probe a running agent โ
cinch collectobserves the real deployment and emits an evidence bundle;cinch verifygrades it. This is the primary path.Review by hand โ
cinch assessgrades a state file a reviewer filled in (or the console session they clicked through).
The grade itself is computed once and shown everywhere: src/cinch/assess.py for the CLI and CI, the dashboard for the browser, both reading their rules โ severity weights, grade bands, critical-gap and completeness caps, remediation phases โ from docs-site/data/rubric.json. Change a rule there and both follow.
# review by hand
cinch assess --state examples/assessment-state.json # grade + insights + action plan
cinch assess --state examples/assessment-state.json --out pack.json
cinch console --assessment pack.json # open the dashboard on it
cinch assess --state pack.json --fail-on critical # gate a release (exit 2)A state file is just what a reviewer decided (see examples/assessment-state.json):
{
"deployment": "checkout-agent-prod",
"status": { "AC-001": "pass", "AC-006": "fail", "AC-014": "na" },
"evidence": { "AC-006": "retrieval output concatenated into the system turn" }
}๐ Assess a running agent
Nobody wants to hand-answer 121 controls. cinch collect observes the deployment and answers what it can prove, in the agent's own environment.
# 1. probe the agent โ out of band, as an operator/sidecar/CI identity
cinch collect --host --pid $AGENT_PID \
--project /srv/checkout-agent \
--deployment checkout-agent-prod \
--sign-cmd 'cosign sign-blob -' \
--out evidence.json
# 2. grade the evidence
cinch verify --evidence evidence.json --out pack.json --fail-on critical
# 3. look at it
cinch console --assessment pack.jsonWhat each probe family observes
Family | Flag | Reads | Controls |
Host |
| The kernel's view of the agent process: UID and |
|
Project |
| The authority the deployment grants: MCP tool grants, container/pod security context, NetworkPolicy egress, committed secret material, dependency pinning and hashes, SBOM, vulnerability scanning, provenance attestation and image signing, CI runner posture, threat model and IR docs |
|
Behaviour |
| How the running agent responds to adversarial input: direct injection, injection smuggled through retrieved content, system-prompt extraction, tool enumeration, rate-limit presence |
|
Host findings are also mapped onto the containment controls that assert the same thing (AE-005 egress โ AC-019 exfiltration), labelled derived: so the chain stays visible.
Probes also abstain when the target is the wrong kind of thing: supply-chain and documentation controls (SBOM, dependency scanning, provenance, threat model) report unknown unless the directory is actually a build tree โ a build manifest, a CI workflow, or a repository. Reporting "no SBOM" about a config directory would be confidently wrong.
Three statuses, and why unknown matters
pass demonstrably enforced ยท fail demonstrably not ยท unknown the probe could not tell.
unknown is never a pass โ and never a fail either. A probe on macOS with no /proc, without CAP_NET_ADMIN, or with no manifest to read says so and leaves the control unreviewed โ it costs completeness instead of quietly earning credit. Keyword hits in config are reported as leads to confirm, not as proof a limit is enforced at runtime. Every observation keeps the raw evidence it was derived from, so you can check the probe's reasoning instead of trusting it. Secret names are recorded; secret values never are.
Self-audit is detected, not assumed away
protocols/evidence-collect.md is blunt about this: an agent auditing its own host is both auditor and audited, and a compromised or simply mistaken agent can report every control as enforced.
So collection records who collected what, where and when โ and when the collector is the audited party, the bundle is stamped provenance.self_attested: true. Three ways that is detected: the inspected PID is the collector's own; the inspected PID is an ancestor of the collector (the agent spawned it, so it inherits the agent's identity and authority); or the request arrived over MCP, where the requester is an agent by definition. That becomes a critical finding in the assessment, is shown as a red self-attested banner in the console, and is never counted as assurance. Independent evidence means cinch collect run out of band under its own identity, signed with --sign-cmd using a key the agent cannot reach.
Behavioural probing sends adversarial input to a live system, so it refuses to run without --authorized naming the target.
As MCP tools
An auditing agent can drive the same flow:
Tool | Use it to |
| Probe a deployment ( |
| Grade a bundle โ score, grade, insights, recommendations, action plan |
Bundles collected through MCP are marked self-attested by construction โ the requester is an agent. Use them to find problems fast; use an out-of-band signed run when the grade has to mean something.
Three ways to get results into the dashboard:
Route | How |
Published pack |
|
Served pack |
|
Loaded by hand | Load assessment in the console header โ accepts an exported pack or a plain state file; |
Exports round-trip: an exported pack feeds straight back into cinch assess --state or the Load assessment button, so re-assessments diff cleanly.
How the grade works. The score is severity-weighted (critical ร3, high ร2, medium/low ร1) over the controls you actually answered. Grade bands run AโF, and two kinds of cap can only ever lower the result โ the pack always says which cap fired and why:
Cap | Effect |
1 unenforced critical control | grade capped at D |
3 or more critical controls | grade capped at F |
under 60% completeness | grade capped at C โ a high score describes a sample, not the deployment |
under 25% completeness | grade I ยท Insufficient evidence โ no letter is claimed at all |
That last one exists because of a real failure: probing a live Claude Code harness resolved 3 of 121 controls, all passing, and the first version graded it A (Contained). A score computed over three controls says nothing about the other 118. Coverage is part of the verdict, not a footnote.
Serve it yourself
The console reads its catalog with fetch('data/full.json'), so it must be served over HTTP โ opening the file directly with file:// makes that fetch fail and falls back to a small offline subset. cinch console does it for you:
pip install cinch-ai-security
cinch console # serves on http://127.0.0.1:8787 and opens your browsercinch console --port 9000 # pick another port
cinch console --no-browser # headless / remote-friendly
cinch console --host 0.0.0.0 # expose on your LAN (read-only, but consider who can reach it)The server is read-only, GET/HEAD only, bound to 127.0.0.1 by default, and rooted at the console directory. Assessment state lives in the browser only โ nothing is uploaded anywhere.
From a source checkout, the same page is served straight from docs-site/:
git clone https://github.com/nirhalfon/cinch-ai-security.git
cd cinch-ai-security && pip install -e .
python scripts/build_docs_json.py # regenerate the catalog bundle after editing any YAML
cinch consoleTurn a session into evidence
Assess the controls that apply to your deployment (Control library โ expand โ Enforced / Gap / N-A, plus an evidence note).
Read Insights & plan โ the grade, what is actually wrong, and what to do in what order.
Export evidence โ Download all โ three artifacts:
JSON โ the full result pack: summary + grade, coverage, insights, ranked recommendations, action plan, and every control with status and evidence. Commit it, diff the next run, gate CI with
cinch assess --fail-on critical.CSV โ the gap register, one row per control, with remediation phase and rank, for tickets and trackers.
PDF โ printable report: grade and why it was capped, insights, action plan, recommendations, CUSTODY coverage, open gaps, NIST AI RMF crosswalk.
๐ Quick start for AI agents
You are an AI agent reading this repo. This section is for you. Three commands and one config block get you operational.
1. Install
pip install cinch-ai-security
# or, from source:
git clone https://github.com/nirhalfon/cinch-ai-security.git
cd cinch-ai-security && pip install -e .2. Wire it into your MCP client
Add this to your MCP config (Claude Desktop, Claude Code, Cursor, etc.):
{
"mcpServers": {
"cinch": {
"command": "cinch",
"args": ["serve"]
}
}
}3. Query the controls you need
Your tools (all return JSON):
Tool | Use it to |
| Probe a running deployment (host / project / endpoint) and get an evidence bundle |
| Grade an evidence bundle โ score, grade, insights, recommendations, action plan |
| Run a named checklist against a description of your deployment |
| List the 6 checklists and their item counts |
| Get a specific control by ID ( |
| Get a step-by-step protocol by name ( |
| Look up controls mapped to a framework ( |
| Find every control that mitigates a given threat โ scans threat + control + verification + sources |
| Compare two checklists to surface coverage gaps and duplicated controls |
Checklist ID prefixes: AC agent-containment ยท AE agent-environment ยท HE harness-engineering ยท RT red-team ยท SC supply-chain ยท SH system-hardening.
30-second smoke test
cinch serve & # start the MCP server on stdio
# then call checklist_list from your MCP client, or:
python -c "from cinch.loader import list_checklists; [print(c['name'], c['item_count']) for c in list_checklists()]"CLI reference
Command | What it does |
| Run the MCP server over stdio (the default with no subcommand) |
| Serve the assessment console on localhost ( |
| Probe a running agent โ evidence bundle ( |
| Grade an evidence bundle ( |
| Grade a hand-reviewed state file ( |
Use as a skill (Hermes / Claude / OpenClaw / NanoClaw)
# Hermes skill
cp -r skills/ai-harness-review ~/.hermes/skills/
# Claude Code โ drop into your project root; it auto-loads
cp cross-harness/claude/CLAUDE.md /your/project/CLAUDE.md
# Codex / OpenCode โ drop into your project root; both read AGENTS.md
cp cross-harness/codex/AGENTS.md /your/project/AGENTS.md
cp cross-harness/opencode/AGENTS.md /your/project/AGENTS.md
# Gemini CLI โ GEMINI.md in your project root
cp cross-harness/gemini/GEMINI.md /your/project/GEMINI.md
# OpenClaw / NanoClaw โ see cross-harness/openclaw/ and cross-harness/nanoclaw/๐งญ Philosophy
The model proposes; the architecture authorizes and enforces. Prompts are not a security boundary.
An AI agent can be manipulated, compromised, or wrong. Its environment must prevent a bad decision from becoming an unrestricted system action.
Functional correctness โ security. 47.5% of AI-generated code may be functionally correct, but only 8.25% is also secure (ICLR 2026 vibe-coding benchmark).
Capability accretion is the core risk. Agents silently gain practical authority through inherited credentials, trust relationships, tool access, and delegation chains.
Defense in depth maps to architecture layers. A control at one layer does not detect an attack at another (LASM principle).
๐ Project structure
cinch/
โโโ src/cinch/ # MCP server (SDK 2.0)
โ โโโ server.py # 9 tool definitions + CLI (serve / console / collect / verify / assess)
โ โโโ probes/ # observe a running agent
โ โ โโโ host.py # AE-001..AE-011 from /proc, cgroups, netfilter, mountinfo
โ โ โโโ project.py # MCP grants, manifests, CI, pinning, secrets (HE/SC/AC)
โ โ โโโ behaviour.py # injection / leakage / tool / rate probes (authorization-gated)
โ โโโ collect.py # evidence bundle: provenance, self-audit detection, signing
โ โโโ verify.py # bundle โ graded assessment (unknown never becomes a pass)
โ โโโ assess.py # assessment engine: score, grade, insights, recommendations, plan
โ โโโ console.py # localhost static server for the assessment console
โ โโโ loader.py # YAML checklist/protocol/mapping loader (path-traversal-safe)
โโโ checklists/ # 6 checklists ยท 121 controls
โ โโโ agent-containment.yaml # CUSTODY-based agent containment (AC)
โ โโโ agent-environment.yaml # host/container controls, out-of-band audited (AE)
โ โโโ harness-engineering.yaml # AI dev harness safeguards (HE)
โ โโโ system-hardening.yaml # OS/infra hardening for AI workloads (SH)
โ โโโ red-team.yaml # AI red team engagement checklist (RT)
โ โโโ supply-chain.yaml # AI supply chain security (SC)
โโโ protocols/ # 5 how-to procedures
โ โโโ agent-deployment.md
โ โโโ evidence-collect.md # out-of-band signed-JSON host evidence for AE controls
โ โโโ incident-response.md
โ โโโ red-team-engagement.md
โ โโโ harness-setup.md
โโโ skills/ # 3 Hermes skills
โ โโโ ai-harness-review/SKILL.md
โ โโโ agent-audit/SKILL.md # + separation-of-duties / --self-audit rule
โ โโโ ai-red-team/SKILL.md
โโโ mappings/ # 6 framework crosswalks
โ โโโ nist-rmf-crosswalk.yaml
โ โโโ owasp-llm-crosswalk.yaml
โ โโโ owasp-agentic-crosswalk.yaml
โ โโโ atlas-crosswalk.yaml
โ โโโ custody-crosswalk.yaml
โ โโโ lasm-crosswalk.yaml
โโโ cross-harness/ # agent platform configs
โ โโโ claude/CLAUDE.md
โ โโโ codex/AGENTS.md
โ โโโ gemini/GEMINI.md
โ โโโ opencode/AGENTS.md
โ โโโ openclaw/agent-config.yaml
โ โโโ nanoclaw/agent-config.yaml
โโโ templates/ # 4 review/policy/ADR templates
โโโ docs/ # threat-model, research-references, design-rationale
โโโ marketing/ # announcement copy (kept out of docs/ โ docs/ is bundled)
โโโ docs-site/ # the console
โ โโโ index.html # landing page + assessment dashboard (one file)
โ โโโ data/
โ โโโ full.json # generated control catalog
โ โโโ rubric.json # scoring rules โ shared with src/cinch/assess.py
โ โโโ assessment.json # generated result pack the console opens on
โโโ examples/assessment-state.json # worked assessment state (input to `cinch assess`)
โโโ scripts/
โ โโโ build_docs_json.py # deterministic full.json generator (run on every source change)
โ โโโ build_assessment.py # regenerate docs-site/data/assessment.json
โ โโโ migrate_schema.py # one-time Schema B โ canonical Schema A migration
โโโ tests/ # pytest: loader, server, console, assessment engine, probes
โโโ .github/
โโโ dependabot.yml # weekly pip + github-actions updates
โโโ workflows/
โโโ validate.yaml # CI: YAML, imports, ruff, bandit, pip-audit, content gate, pytest, full.json sync
โโโ security-scan.yml # CodeQL, semgrep, dependency-review, gitleaks, SBOM
โโโ pages.yml # deploy the console to GitHub Pages๐ฌ Research grounding
Source | Coverage |
Risk governance (GOVERN, MAP, MEASURE, MANAGE) | |
Generative AI-specific risks and controls | |
Secure AI lifecycle controls | |
LLM application security risks | |
Agent threat modeling and controls | |
Adversarial threat landscape for AI | |
Autonomous agent containment | |
Layered Attack Surface Model | |
Secure AI Framework | |
AI Controls Matrix | |
AI cybersecurity standardization | |
Engineering safeguards for AI-assisted dev | |
CI/CD practices, stability metrics | |
Subtle correctness in AI-generated code |
๐ค Contributing
See CONTRIBUTING.md. Every checklist item, protocol step, and mapping must cite a source. After any source change, regenerate both generated files โ CI verifies they are in sync:
python scripts/build_docs_json.py # docs-site/data/full.json (control catalog)
python scripts/build_assessment.py # docs-site/data/assessment.json (result pack the console opens on)๐ Security
See SECURITY.md for vulnerability reporting. This project dogfoods its own controls โ the repo itself is scanned with CodeQL, semgrep, bandit, pip-audit, and gitleaks on every push.
๐ License
CC BY 4.0 โ free to share and adapt with attribution.
Available Tools
9 toolschecklist_diffA
Compare two checklists and show coverage gaps and duplicated controls.
Returns items present in checklist A but missing from B, and vice versa. Matching is by normalized control text (case- and whitespace-insensitive), not by item ID, so the same control stated in two checklists (e.g. agent-containment and harness-engineering) is recognized as a duplicate. Use this to keep the catalog maintainable as it grows.
Args: a: First checklist name, e.g. 'agent-containment'. b: Second checklist name, e.g. 'harness-engineering'.
Inputs: a (required), b (required). Returns: result
checklist_getA
Get a specific checklist item by its ID (e.g. 'AC-001', 'HE-006').
Returns the full item with threat, control, severity, verification, and source framework references.
Args: checklist_name: Checklist name, e.g. 'agent-containment'. item_id: Item ID, e.g. 'AC-001'.
Inputs: item_id (required), checklist_name (required). Returns: result
checklist_listA
List all available security checklists with item counts and frameworks.
Returns checklist name, version, description, item count, and mapped frameworks. No arguments.
Inputs: no named parameters. Returns: result
checklist_runA
Run a named checklist against a description of your AI agent or deployment.
Returns all items in the checklist with their threat, control, severity, and verification steps. Use this to assess whether your deployment meets security controls from NIST, CISA, OWASP, CUSTODY, and LASM.
Args: name: Checklist name, e.g. 'agent-containment', 'harness-engineering', 'system-hardening', 'red-team', or 'supply-chain'.
Inputs: name (required). Returns: result
evidence_collectA
Probe a running agent deployment and return an evidence bundle.
Runs read-only probes and returns what was actually observed, per control: 'pass' (demonstrably enforced), 'fail' (demonstrably not), 'unknown' (could not tell โ never counted as enforced). Feed the bundle to evidence_verify to get a grade, insights and an action plan.
Self-audit warning: when an agent calls this tool about its own host, the result
is self-attestation, not assurance. The bundle records that ('provenance.
self_attested') and the assessment raises it as a critical finding. Independent
evidence needs cinch collect run out of band under a separate identity โ see
the 'evidence-collect' protocol.
Args:
host: Probe the host/container running the agent (AE-001..AE-011).
pid: PID of the agent process to inspect. Omit and the collector inspects
itself, which is flagged as self-attestation.
unit: systemd unit name of the agent, used to resolve its MainPID.
project_path: Deployment directory to inspect โ MCP tool grants, container
manifests, CI workflows, dependency pinning, secret handling.
endpoint: Running agent's HTTP endpoint for behavioural probes (prompt
injection, prompt leakage, tool enumeration, rate bounds).
authorized: Must be true to probe endpoint. Confirms you are permitted to
send adversarial input to that target.
deployment: Name recorded in the bundle.
Inputs: pid, host, unit, endpoint, +3 more. Returns: result
evidence_verifyA
Grade an evidence bundle: score, letter grade, insights, recommendations, plan.
Applies the same rubric a human reviewer sees in the console, and reports on the evidence itself as well as the controls โ self-attested collection, unsigned bundles, and controls no probe could verify all surface as findings. 'unknown' observations stay unreviewed rather than counting as enforced.
Args: bundle_json: A 'cinch-evidence/1' bundle, as returned by evidence_collect. deployment: Optional deployment name override.
Inputs: bundle_json (required), deployment. Returns: result
mapping_lookupA
Look up security controls mapped to a specific framework.
Returns all checklist items that correspond to framework controls, enabling cross-reference between NIST AI RMF, OWASP, CUSTODY, LASM, and MITRE ATLAS.
Args: framework: Framework name: 'nist-rmf', 'owasp-llm', 'atlas', 'custody', or 'lasm'.
Inputs: framework (required). Returns: result
protocol_getA
Get a step-by-step security protocol by name.
Returns the full protocol document with prerequisites, steps, verification, and rollback procedures.
Args: name: Protocol name, e.g. 'agent-deployment', 'incident-response', 'red-team-engagement', or 'harness-setup'.
Inputs: name (required). Returns: result
threat_searchA
Search all checklists for controls that mitigate a given threat.
Use this to find relevant controls for specific threats like 'prompt injection', 'credential theft', 'data exfiltration', 'privilege escalation', or 'lateral movement'.
Args: query: Threat or control to search for.
Inputs: query (required). Returns: result
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelogโฆ
TDQS
Scored across 9 tools
Each tool targets a distinct operation: listing, running, retrieving, diffing checklists; retrieving protocols; mapping frameworks; searching threats; and collecting/verifying evidence. Even closely related tools like checklist_run and evidence_collect are clearly separated by static assessment vs. live probing.
All tools use a consistent verb_noun snake_case pattern (e.g., checklist_list, protocol_get, evidence_verify). Minor synonym use like 'lookup' and 'search' does not break the overall predictable style.
Nine tools is well-scoped for a security checklist/evidence server, covering discovery, assessment, cross-referencing, and evidence workflows without redundancy or bloat.
The set covers the core lifecycle: listing/running/retrieving/checklists, fetching protocols, mapping to frameworks, searching threats, and collecting/verifying evidence. The only notable gap is the lack of a protocol_list endpoint to enumerate available protocols without knowing names in advance, but this is workable.
Maintenance
Related MCP Connectors
Find, vet, and run MCP tools through a secure audited gateway with prompt-injection risk scoring
AI-security knowledge as MCP: standards-mapped tools (OWASP, NIST, MITRE) for AI agents.
Compliance frameworks (SOC 2, ISO 27001, CMMC, NIST, more) delivered to AI agents as MCP tools.
Read-only search of 2,189 AI agent tools (MCP servers, x402 APIs) with daily automated trust scores.
Related MCP Servers
- FlicenseAqualityDmaintenanceEnables AI assistants to perform defensive security tasks such as vulnerability detection, CVE lookup, phishing/link safety checks, and security report generation via MCP tools.23-
- AlicenseNot gradedqualityBmaintenanceEnables AI agents to orchestrate 100+ security tools over MCP for authorized penetration testing, including recon, scanning, exploitation, attack-chain planning, and knowledge-base retrieval.7Apache 2.0
- AlicenseNot gradedqualityDmaintenanceEnables AI agents to autonomously run over 150 cybersecurity tools for penetration testing, vulnerability discovery, bug bounty automation, and security research through MCP.MIT
- AlicenseNot gradedqualityCmaintenanceEnables AI agents to connect over MCP and orchestrate 150+ security CLI tools and autonomous agents for reconnaissance, penetration testing, vulnerability analysis, and reporting.MIT