Please report security issues privately via GitHub Security Advisories rather than a public issue. Include a description of the issue, steps to reproduce, and the affected version.
ComputeLedger's security model is entirely local: signing keys, the ledger, and all writes stay on the machine running the CLI. There is no server component, no account system, and no network calls other than what the mcp subcommand's stdio transport requires from whatever MCP client invokes it.
Relevant areas for a security report:
- Signature forgery or verification bypass (
src/receipt.ts,src/crypto.ts,python/src/computeledger/receipt.py,python/src/computeledger/crypto.py) - Hash-chain tampering that
ledger verifyfails to detect (src/ledger.ts,python/src/computeledger/ledger.py) - Command injection via
computeledger run(src/run.ts,python/src/computeledger/run.py) - Key storage or file-permission issues (
src/crypto.ts,python/src/computeledger/crypto.py) - Cross-language interoperability bugs that make a legitimately signed receipt fail verification in the other implementation, or an invalid receipt pass
Only the latest published release on npm and PyPI receives security fixes.