Chitin

Armor for your AI agents

53% of OpenClaw skills are unsafe. Chitin adds a security layer with a verified skill marketplace, runtime enforcement, and a global revocation network.

$openclaw plugins install @chitin/openclaw

0

Skills Scanned

0

Threats Blocked

0

Revocations Issued

Security that works with you

We scan. We enforce. We revoke. All without slowing you down.

Verified Marketplace

Browse reviewed skill versions with Chitin scan results for prompt injection, data exfiltration, and related risks.

Our pipeline combines static analysis, Semgrep rules, SBOM/CVE checks, and LLM security review for the reviewed artifact.

Runtime Enforcement

Block revoked skills and policy violations before tool calls run. The Chitin plugin checks installs and tool calls against your configured permissions.

Fail-closed by default — if something goes wrong, the tool call is blocked, not allowed. Per-skill permission scopes with domain allowlisting.

Global Revocation Network

When Chitin revokes a skill, protected agents receive updated revocation data on the next sync, typically within 30 seconds.

Revocation data helps users respond quickly as new information is discovered. Depending on user policy, revoked skills are blocked and may be removed.

Verified Marketplace

Browse reviewed skill versions with Chitin scan results for prompt injection, data exfiltration, and related risks.

web-search
95
file-manager
88
github-int
91
slack-notify
93

Verified skills, ready to install

Runtime Enforcement

Block revoked skills and policy violations before tool calls run. The Chitin plugin checks installs and tool calls against your configured permissions.

$ openclaw skills install web-search

✓ Chitin: Verified (score: 95)

$ openclaw skills install super-helper-pro

✗ Chitin: REVOKED — data exfiltration detected

$ openclaw skills install weather-fetcher

⚠ Chitin: Unverified — proceed with caution

$ openclaw skills install code-runner

✓ Chitin: Verified (score: 82)

Global Revocation Network

When Chitin revokes a skill, protected agents receive updated revocation data on the next sync, typically within 30 seconds.

Monitoring network...

OWASP Top 10 for Agentic Applications

Chitin maps concrete controls to all 10 categories in OWASP's peer-reviewed Top 10 for Agentic Applications (2026).

ASI01Covered

Agent Goal Hijack

Prompt-injection detection, Unicode evasion handling, and counter-prompts help prevent external content from redirecting agent goals.

ASI02Covered

Tool Misuse & Exploitation

Runtime permission checks, domain allowlists, C2 blocking, and sensitive-path guards constrain unsafe tool use.

ASI03Covered

Identity & Privilege Abuse

Per-skill scoped permissions, fail-closed hooks, and API/session scoping limit privilege abuse.

ASI04Covered

Agentic Supply Chain Vulnerabilities

Static analysis, SBOM extraction, suspicious dependency checks, OSV CVE lookup, and revocation status checks cover skill supply-chain risk.

ASI05Covered

Unexpected Code Execution (RCE)

Interpreter mapping, shell-scope enforcement, dangerous command detection, and Semgrep rules surface unexpected execution patterns.

ASI06Covered

Memory & Context Poisoning

Runtime guards block writes to Chitin state, cognitive files, and sensitive paths; prompt/context tampering patterns are detected.

ASI07Covered

Insecure Inter-Agent Communication

Subagent-spawn authorization and lateral traffic scanners cover explicit and content-mediated inter-agent communication risks.

ASI08Covered

Cascading Failures

Ancestor-aware rate limiting charges child calls to parent skills, containing fanout loops and cascading failures.

ASI09Covered

Human-Agent Trust Exploitation

Social-engineering detectors and separate counter-prompts flag manipulation, credential harvesting, and unsafe human-in-the-loop pressure.

ASI10Covered

Rogue Agents

Verified-skill checks, runtime policy enforcement, suspicious-behavior signals, and global revocation help contain compromised or divergent agents.

Protect your agents today

Free to start. Upgrade anytime for continuous revocation monitoring and the security dashboard.