More screenshots

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    cipherhub crypto

    1

    Production-grade MCP cryptography toolkit for hashing, encryption, PQC migration, and X.509/ASN.1 parsing.

    Free

    1 useSecurity scanned
    cipherhub crypto

    cipherhub crypto

    Example session with this skill installed

    I'm finalizing firmware release v2.4.1 for our OTA update channel. The release header is the ASCII string CIPHERHUB-FW-2.4.1 (hex: 4349504845524855422d46572d322e342e31). Compute SHA-384 and SM3 digests so I can publish the checksums alongside the image. Then walk me through a complete X25519 + ML-KEM-768 hybrid key exchange — both sides, every intermediate value — so I can see exactly what the post-quantum handshake for the update channel looks like.

    • Read your context and instructions
    • Compiled the cipherhub crypto

    Summary

    Two digest computations and one full post-quantum key exchange, all executed live against the CipherHUB MCP service. The hybrid handshake completed with matching session keys on both sides (keys_match: true).

    Release checksums — CIPHERHUB-FW-2.4.1

    Input (hex): 4349504845524855422d46572d322e342e31

    AlgorithmDigestBytes
    SHA-384e3b47093240eb98cbd69d3a8b5e360310f0782fea2ae0e5fae5ea5da860d92b79b828cbf89c33adbd008084eefb84a6748
    SM3e3e34e92a93ffb7a8faf7407ec0253dd12ff1d5726accf225fe9d030215a163832

    Both digests were computed in a single batched hash_sum call.

    Hybrid X25519 + ML-KEM-768 exchange

    The hybrid_kex tool ran the full handshake and returned every intermediate value:

    StepResult
    1. X25519 ECDHAlice and Bob derive the same shared secret 0b6c0a4f…33ed374
    2. ML-KEM-768 encapsulation1088-byte ciphertext; both sides recover secret 8f1021b7…a2b74727
    3. CombinerIKM = ECDH secret ‖ KEM secret; HKDF-SHA256, salt X-Wing-KEM-v1 (X-Wing construction)
    4. Final session keyfea3be84e84abd69797b3e69f19b7ba1d837d9ce41c830439ad41c81bfdcdf19 — identical on both sides

    If a quantum adversary breaks X25519, the ML-KEM half still holds the channel; if ML-KEM is broken, classical ECDH does. The transcript also includes PEM key pairs and the full ciphertext hex for offline inspection.

    Next steps

    • Publish the SHA-384 digest with the release; clients verify after download
    • Re-run hybrid_kex per session — the service is stateless, nothing is retained
    • Same service also covers SM2/SM9, ML-DSA signatures, certificates, COSE/CMS — 55 tools in total

    Connects securely to your tools. The creator never sees your data.

    About this skill

    Handling raw binary data and complex cryptographic operations is error-prone for AI agents. This skill provides a production-grade interface to the CipherHUB MCP server, enabling secure hashing, encryption, key management, and post-quantum operations without managing local libraries.

    What it does

    • Multi-algorithm hashing provides SHA-2 family, SM3, and SHAKE digests with support for large inputs up to 16MB.
    • Advanced encryption handles AES (CBC, GCM, CTR, XTS), ChaCha20-Poly1305, and Chinese national standard SM4/ZUC ciphers.
    • Asymmetric key management generates and processes RSA, ECC (NIST and Bernstein curves), and SM2/SM9 identity-based keys.
    • Post-quantum security implements FIPS 203 (ML-KEM) and FIPS 204 (ML-DSA) for future-proof migrations.
    • Structured data parsing dissects ASN.1/DER, CBOR, X.509 certificates, and COSE/CMS signed or enveloped messages.
    • Cryptographic primitives generates CSPRNG random data and performs HKDF-based key derivation.

    How it works

    1. Hex encoding converts all binary input data into even-length hex strings before transmission.
    2. Tool selection identifies the specific algorithm or protocol operation from the 55 available tools.
    3. JSON-RPC execution sends a stateless POST request to the hosted CipherHUB endpoint or local MCP runtime.
    4. Response handling parses the double-wrapped JSON output and decodes the result hex strings back into the target format.

    Frameworks & tools

    Built for the Model Context Protocol (MCP) using JSON-RPC 2.0. Verified for interoperability with AWS KMS, Tencent Cloud KMS, and OpenSSL 3.6.

    Why this beats prompting it yourself

    LLMs frequently hallucinate binary formats, padding schemes, and ASN.1 structures. This skill offloads the heavy lifting to a specialized GmSSL-backed service, ensuring mathematical correctness and protocol compliance that raw prompting cannot achieve.

    Use cases

    • Generating secure X.509 certificate signing requests or parsing existing chains for validation.
    • Implementing hybrid post-quantum key exchanges (X25519 + ML-KEM-768) in agent workflows.
    • Verifying the integrity of large files using batch hash_sum operations.
    • Encrypting sensitive agent state or logs using AES-GCM or SM4-GCM.

    Known limitations

    Inputs are limited to 16MB for hashing and 128KB for structured ASN.1/CBOR parsing. Requests are rate-limited to 10 per second with a burst of 20.

    Hex strings must be even-length. Hashing capped at 16MB. Random data limited to 128 bytes per call. ASN.1/CBOR parsing capped at 128KB. Rate limited to 10 req/s.

    How to use

    1. 1

      Connect your agent to the Agensi MCP

      Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup

    2. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before any payment.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill goes through the trust checks below.

    Security scanned

    Checked by our 8-point scan before listing

    1 use

    Times used to date

    Free forever

    No account required to browse

    Trust & safety

    Security scan

    Security scan passed 1 month ago

    • Free with an Agensi account

    Requires

    Cipherhub

    Permissions

    Terminal / Shell
    Network Access

    Allowed Hosts

    tools.cipherhub.cloud
    cipherhub.cloud
    oidc.cipherhub.cloud

    Network Access is the only hard requirement: the skill calls the hosted CipherHUB cryptography service at tools.cipherhub.cloud over plain HTTPS (JSON-RPC POST). Terminal/Shell is used for the documented curl examples and the optional `claude mcp add` registration command. The skill reads no files, writes no files, uses no environment variables, and requires no accounts or API keys.

    Listed1 month ago

    What's inside

    Frequently Asked Questions

    This skill allows AI agents to perform complex binary operations like PQC (Post-Quantum Cryptography) and ASN.1 parsing, which LLMs typically fail at due to 'hallucinations' in bitwise logic and padding. It ensures mathematical correctness by offloading these tasks to a production-grade GmSSL-backed engine.

    The skill is built on the Model Context Protocol (MCP) and is compatible with any agentic framework that supports MCP, including Claude Desktop, LangChain, and various autonomous agent IDEs. It also integrates seamlessly with cloud environments like AWS and Tencent Cloud KMS.

    All binary inputs, such as keys or ciphertext, must be converted into even-length hex strings before being passed to the tool. The skill then returns results in hex format, which your agent can easily decode or pass to subsequent cryptographic operations.

    CipherHUB Crypto includes full support for FIPS 203 (ML-KEM) and FIPS 204 (ML-DSA). This allows you to build future-proof agents capable of generating post-quantum signatures and performing hybrid key exchanges today.

    You receive access to a suite of 55 specialized tools covering hashing, symmetric/asymmetric encryption, and structured data parsing. This includes support for global standards like AES and RSA, as well as Chinese national standards like SM2, SM3, and SM4.

    Hashing operations support inputs up to 16MB, while structured parsing for ASN.1, CBOR, and X.509 certificates is limited to 128KB per request. This ensures high-performance processing while maintaining stability for the MCP runtime.

    New in Security